Datasets:

CShorten

/

Last-Week-on-ML-ArXiv

like 0

What's the Harm? Sharp Bounds on the Fraction Negatively Affected by Treatment

The fundamental problem of causal inference -- that we never observe counterfactuals -- prevents us from identifying how many might be negatively affected by a proposed intervention. If, in an A/B test, half of users click (or buy, or watch, or renew, etc.), whether exposed to the standard experience A or a new one B, hypothetically it could be because the change affects no one, because the change positively affects half the user population to go from no-click to click while negatively affecting the other half, or something in between. While unknowable, this impact is clearly of material importance to the decision to implement a change or not, whether due to fairness, long-term, systemic, or operational considerations. We therefore derive the tightest-possible (i.e., sharp) bounds on the fraction negatively affected (and other related estimands) given data with only factual observations, whether experimental or observational. Naturally, the more we can stratify individuals by observable covariates, the tighter the sharp bounds. Since these bounds involve unknown functions that must be learned from data, we develop a robust inference algorithm that is efficient almost regardless of how and how fast these functions are learned, remains consistent when some are mislearned, and still gives valid conservative bounds when most are mislearned. Our methodology altogether therefore strongly supports credible conclusions: it avoids spuriously point-identifying this unknowable impact, focusing on the best bounds instead, and it permits exceedingly robust inference on these. We demonstrate our method in simulation studies and in a case study of career counseling for the unemployed.

Nothing makes sense in deep learning, except in the light of evolution

Deep Learning (DL) is a surprisingly successful branch of machine learning. The success of DL is usually explained by focusing analysis on a particular recent algorithm and its traits. Instead, we propose that an explanation of the success of DL must look at the population of all algorithms in the field and how they have evolved over time. We argue that cultural evolution is a useful framework to explain the success of DL. In analogy to biology, we use `development' to mean the process converting the pseudocode or text description of an algorithm into a fully trained model. This includes writing the programming code, compiling and running the program, and training the model. If all parts of the process don't align well then the resultant model will be useless (if the code runs at all!). This is a constraint. A core component of evolutionary developmental biology is the concept of deconstraints -- these are modification to the developmental process that avoid complete failure by automatically accommodating changes in other components. We suggest that many important innovations in DL, from neural networks themselves to hyperparameter optimization and AutoGrad, can be seen as developmental deconstraints. These deconstraints can be very helpful to both the particular algorithm in how it handles challenges in implementation and the overall field of DL in how easy it is for new ideas to be generated. We highlight how our perspective can both advance DL and lead to new insights for evolutionary biology.

Seeking entropy: complex behavior from intrinsic motivation to occupy action-state path space

Intrinsic motivation generates behaviors that do not necessarily lead to immediate reward, but help exploration and learning. Here we show that agents having the sole goal of maximizing occupancy of future actions and states, that is, moving and exploring on the long term, are capable of complex behavior without any reference to external rewards. We find that action-state path entropy is the only measure consistent with additivity and other intuitive properties of expected future action-state path occupancy. We provide analytical expressions that relate the optimal policy with the optimal state-value function, from where we prove uniqueness of the solution of the associated Bellman equation and convergence of our algorithm to the optimal state-value function. Using discrete and continuous state tasks, we show that `dancing', hide-and-seek and a basic form of altruistic behavior naturally result from entropy seeking without external rewards. Intrinsically motivated agents can objectively determine what states constitute rewards, exploiting them to ultimately maximize action-state path entropy.

ClusterEA: Scalable Entity Alignment with Stochastic Training and Normalized Mini-batch Similarities

Entity alignment (EA) aims at finding equivalent entities in different knowledge graphs (KGs). Embedding-based approaches have dominated the EA task in recent years. Those methods face problems that come from the geometric properties of embedding vectors, including hubness and isolation. To solve these geometric problems, many normalization approaches have been adopted for EA. However, the increasing scale of KGs renders it hard for EA models to adopt the normalization processes, thus limiting their usage in real-world applications. To tackle this challenge, we present ClusterEA, a general framework that is capable of scaling up EA models and enhancing their results by leveraging normalization methods on mini-batches with a high entity equivalent rate. ClusterEA contains three components to align entities between large-scale KGs, including stochastic training, ClusterSampler, and SparseFusion. It first trains a large-scale Siamese GNN for EA in a stochastic fashion to produce entity embeddings. Based on the embeddings, a novel ClusterSampler strategy is proposed for sampling highly overlapped mini-batches. Finally, ClusterEA incorporates SparseFusion, which normalizes local and global similarity and then fuses all similarity matrices to obtain the final similarity matrix. Extensive experiments with real-life datasets on EA benchmarks offer insight into the proposed framework, and suggest that it is capable of outperforming the state-of-the-art scalable EA framework by up to 8 times in terms of Hits@1.

Delator: Automatic Detection of Money Laundering Evidence on Transaction Graphs via Neural Networks

Money laundering is one of the most relevant criminal activities today, due to its potential to cause massive financial losses to governments, banks, etc. We propose DELATOR, a new CAAT (computer-assisted audit technology) to detect money laundering activities based on neural network models that encode bank transfers as a large-scale temporal graph. In collaboration with a Brazilian bank, we design and apply an evaluation strategy to quantify DELATOR's performance on historic data comprising millions of clients. DELATOR outperforms an off-the-shelf solution from Amazon AWS by 18.9% with respect to AUC. We conducted real experiments that led to discovery of 8 new suspicious among 100 analyzed cases, which would have been reported to the authorities under the current criteria.

On the SDEs and Scaling Rules for Adaptive Gradient Algorithms

Approximating Stochastic Gradient Descent (SGD) as a Stochastic Differential Equation (SDE) has allowed researchers to enjoy the benefits of studying a continuous optimization trajectory while carefully preserving the stochasticity of SGD. Analogous study of adaptive gradient methods, such as RMSprop and Adam, has been challenging because there were no rigorously proven SDE approximations for these methods. This paper derives the SDE approximations for RMSprop and Adam, giving theoretical guarantees of their correctness as well as experimental validation of their applicability to common large-scaling vision and language settings. A key practical result is the derivation of a $\textit{square root scaling rule}$ to adjust the optimization hyperparameters of RMSprop and Adam when changing batch size, and its empirical validation in deep learning settings.

Heterformer: A Transformer Architecture for Node Representation Learning on Heterogeneous Text-Rich Networks

We study node representation learning on heterogeneous text-rich networks, where nodes and edges are multi-typed and some types of nodes are associated with text information. Although recent studies on graph neural networks (GNNs) and pretrained language models (PLMs) have demonstrated their power in encoding network and text signals, respectively, less focus has been given to delicately coupling these two types of models on heterogeneous text-rich networks. Specifically, existing GNNs rarely model text in each node in a contextualized way; existing PLMs can hardly be applied to characterize graph structures due to their sequence architecture. In this paper, we propose Heterformer, a Heterogeneous GNN-nested transformer that blends GNNs and PLMs into a unified model. Different from previous "cascaded architectures" that directly add GNN layers upon a PLM, our Heterformer alternately stacks two modules - a graph-attention-based neighbor aggregation module and a transformer-based text and neighbor joint encoding module - to facilitate thorough mutual enhancement between network and text signals. Meanwhile, Heterformer is capable of characterizing network heterogeneity and nodes without text information. Comprehensive experiments on three large-scale datasets from different domains demonstrate the superiority of Heterformer over state-of-the-art baselines in link prediction, transductive/inductive node classification, node clustering, and semantics-based retrieval.

Pre-Train Your Loss: Easy Bayesian Transfer Learning with Informative Priors

Deep learning is increasingly moving towards a transfer learning paradigm whereby large foundation models are fine-tuned on downstream tasks, starting from an initialization learned on the source task. But an initialization contains relatively little information about the source task. Instead, we show that we can learn highly informative posteriors from the source task, through supervised or self-supervised approaches, which then serve as the basis for priors that modify the whole loss surface on the downstream task. This simple modular approach enables significant performance gains and more data-efficient learning on a variety of downstream classification and segmentation tasks, serving as a drop-in replacement for standard pre-training strategies. These highly informative priors also can be saved for future use, similar to pre-trained weights, and stand in contrast to the zero-mean isotropic uninformative priors that are typically used in Bayesian deep learning.

DEMAND: Deep Matrix Approximately Nonlinear Decomposition to Identify Meta, Canonical, and Sub-Spatial Pattern of functional Magnetic Resonance Imaging in the Human Brain

Deep Neural Networks (DNNs) have already become a crucial computational approach to revealing the spatial patterns in the human brain; however, there are three major shortcomings in utilizing DNNs to detect the spatial patterns in functional Magnetic Resonance Signals: 1). It is a fully connected architecture that increases the complexity of network structures that is difficult to optimize and vulnerable to overfitting; 2). The requirement of large training samples results in erasing the individual/minor patterns in feature extraction; 3). The hyperparameters are required to be tuned manually, which is time-consuming. Therefore, we propose a novel deep nonlinear matrix factorization named Deep Matrix Approximately Nonlinear Decomposition (DEMAND) in this work to take advantage of the shallow linear model, e.g., Sparse Dictionary Learning (SDL) and DNNs. At first, the proposed DEMAND employs a non-fully connected and multilayer-stacked architecture that is easier to be optimized compared with canonical DNNs; furthermore, due to the efficient architecture, training DEMAND can avoid overfitting and enables the recognition of individual/minor features based on a small dataset such as an individual data; finally, a novel rank estimator technique is introduced to tune all hyperparameters of DEMAND automatically. Moreover, the proposed DEMAND is validated by four other peer methodologies via real functional Magnetic Resonance Imaging data in the human brain. In short, the validation results demonstrate that DEMAND can reveal the reproducible meta, canonical, and sub-spatial features of the human brain more efficiently than other peer methodologies.

Adaptive Fairness-Aware Online Meta-Learning for Changing Environments

The fairness-aware online learning framework has arisen as a powerful tool for the continual lifelong learning setting. The goal for the learner is to sequentially learn new tasks where they come one after another over time and the learner ensures the statistic parity of the new coming task across different protected sub-populations (e.g. race and gender). A major drawback of existing methods is that they make heavy use of the i.i.d assumption for data and hence provide static regret analysis for the framework. However, low static regret cannot imply a good performance in changing environments where tasks are sampled from heterogeneous distributions. To address the fairness-aware online learning problem in changing environments, in this paper, we first construct a novel regret metric FairSAR by adding long-term fairness constraints onto a strongly adapted loss regret. Furthermore, to determine a good model parameter at each round, we propose a novel adaptive fairness-aware online meta-learning algorithm, namely FairSAOML, which is able to adapt to changing environments in both bias control and model precision. The problem is formulated in the form of a bi-level convex-concave optimization with respect to the model's primal and dual parameters that are associated with the model's accuracy and fairness, respectively. The theoretic analysis provides sub-linear upper bounds for both loss regret and violation of cumulative fairness constraints. Our experimental evaluation on different real-world datasets with settings of changing environments suggests that the proposed FairSAOML significantly outperforms alternatives based on the best prior online learning approaches.

Persistent Homology of Coarse Grained State Space Networks

This work is dedicated to the topological analysis of complex transitional networks for dynamic state detection. Transitional networks are formed from time series data and they leverage graph theory tools to reveal information about the underlying dynamic system. However, traditional tools can fail to summarize the complex topology present in such graphs. In this work, we leverage persistent homology from topological data analysis to study the structure of these networks. We contrast dynamic state detection from time series using CGSSN and TDA to two state of the art approaches: Ordinal Partition Networks (OPNs) combined with TDA, and the standard application of persistent homology to the time-delay embedding of the signal. We show that the CGSSN captures rich information about the dynamic state of the underlying dynamical system as evidenced by a significant improvement in dynamic state detection and noise robustness in comparison to OPNs. We also show that because the computational time of CGSSN is not linearly dependent on the signal's length, it is more computationally efficient than applying TDA to the time-delay embedding of the time series.

Explanatory machine learning for sequential human teaching

The topic of comprehensibility of machine-learned theories has recently drawn increasing attention. Inductive Logic Programming (ILP) uses logic programming to derive logic theories from small data based on abduction and induction techniques. Learned theories are represented in the form of rules as declarative descriptions of obtained knowledge. In earlier work, the authors provided the first evidence of a measurable increase in human comprehension based on machine-learned logic rules for simple classification tasks. In a later study, it was found that the presentation of machine-learned explanations to humans can produce both beneficial and harmful effects in the context of game learning. We continue our investigation of comprehensibility by examining the effects of the ordering of concept presentations on human comprehension. In this work, we examine the explanatory effects of curriculum order and the presence of machine-learned explanations for sequential problem-solving. We show that 1) there exist tasks A and B such that learning A before B has a better human comprehension with respect to learning B before A and 2) there exist tasks A and B such that the presence of explanations when learning A contributes to improved human comprehension when subsequently learning B. We propose a framework for the effects of sequential teaching on comprehension based on an existing definition of comprehensibility and provide evidence for support from data collected in human trials. Empirical results show that sequential teaching of concepts with increasing complexity a) has a beneficial effect on human comprehension and b) leads to human re-discovery of divide-and-conquer problem-solving strategies, and c) studying machine-learned explanations allows adaptations of human problem-solving strategy with better performance.

SADAM: Stochastic Adam, A Stochastic Operator for First-Order Gradient-based Optimizer

In this work, to efficiently help escape the stationary and saddle points, we propose, analyze, and generalize a stochastic strategy performed as an operator for a first-order gradient descent algorithm in order to increase the target accuracy and reduce time consumption. Unlike existing algorithms, the proposed stochastic the strategy does not require any batches and sampling techniques, enabling efficient implementation and maintaining the initial first-order optimizer's convergence rate, but provides an incomparable improvement of target accuracy when optimizing the target functions. In short, the proposed strategy is generalized, applied to Adam, and validated via the decomposition of biomedical signals using Deep Matrix Fitting and another four peer optimizers. The validation results show that the proposed random strategy can be easily generalized for first-order optimizers and efficiently improve the target accuracy.

EXODUS: Stable and Efficient Training of Spiking Neural Networks

Spiking Neural Networks (SNNs) are gaining significant traction in machine learning tasks where energy-efficiency is of utmost importance. Training such networks using the state-of-the-art back-propagation through time (BPTT) is, however, very time-consuming. Previous work by Shrestha and Orchard [2018] employs an efficient GPU-accelerated back-propagation algorithm called SLAYER, which speeds up training considerably. SLAYER, however, does not take into account the neuron reset mechanism while computing the gradients, which we argue to be the source of numerical instability. To counteract this, SLAYER introduces a gradient scale hyperparameter across layers, which needs manual tuning. In this paper, (i) we modify SLAYER and design an algorithm called EXODUS, that accounts for the neuron reset mechanism and applies the Implicit Function Theorem (IFT) to calculate the correct gradients (equivalent to those computed by BPTT), (ii) we eliminate the need for ad-hoc scaling of gradients, thus, reducing the training complexity tremendously, (iii) we demonstrate, via computer simulations, that EXODUS is numerically stable and achieves a comparable or better performance than SLAYER especially in various tasks with SNNs that rely on temporal features. Our code is available at https://github.com/synsense/sinabs-exodus.

Exploring the Trade-off between Plausibility, Change Intensity and Adversarial Power in Counterfactual Explanations using Multi-objective Optimization

There is a broad consensus on the importance of deep learning models in tasks involving complex data. Often, an adequate understanding of these models is required when focusing on the transparency of decisions in human-critical applications. Besides other explainability techniques, trustworthiness can be achieved by using counterfactuals, like the way a human becomes familiar with an unknown process: by understanding the hypothetical circumstances under which the output changes. In this work we argue that automated counterfactual generation should regard several aspects of the produced adversarial instances, not only their adversarial capability. To this end, we present a novel framework for the generation of counterfactual examples which formulates its goal as a multi-objective optimization problem balancing three different objectives: 1) plausibility, i.e., the likeliness of the counterfactual of being possible as per the distribution of the input data; 2) intensity of the changes to the original input; and 3) adversarial power, namely, the variability of the model's output induced by the counterfactual. The framework departs from a target model to be audited and uses a Generative Adversarial Network to model the distribution of input data, together with a multi-objective solver for the discovery of counterfactuals balancing among these objectives. The utility of the framework is showcased over six classification tasks comprising image and three-dimensional data. The experiments verify that the framework unveils counterfactuals that comply with intuition, increasing the trustworthiness of the user, and leading to further insights, such as the detection of bias and data misrepresentation.

Learning Task-relevant Representations for Generalization via Characteristic Functions of Reward Sequence Distributions

Generalization across different environments with the same tasks is critical for successful applications of visual reinforcement learning (RL) in real scenarios. However, visual distractions -- which are common in real scenes -- from high-dimensional observations can be hurtful to the learned representations in visual RL, thus degrading the performance of generalization. To tackle this problem, we propose a novel approach, namely Characteristic Reward Sequence Prediction (CRESP), to extract the task-relevant information by learning reward sequence distributions (RSDs), as the reward signals are task-relevant in RL and invariant to visual distractions. Specifically, to effectively capture the task-relevant information via RSDs, CRESP introduces an auxiliary task -- that is, predicting the characteristic functions of RSDs -- to learn task-relevant representations, because we can well approximate the high-dimensional distributions by leveraging the corresponding characteristic functions. Experiments demonstrate that CRESP significantly improves the performance of generalization on unseen environments, outperforming several state-of-the-arts on DeepMind Control tasks with different visual distractions.

Memorization and Optimization in Deep Neural Networks with Minimum Over-parameterization

The Neural Tangent Kernel (NTK) has emerged as a powerful tool to provide memorization, optimization and generalization guarantees in deep neural networks. A line of work has studied the NTK spectrum for two-layer and deep networks with at least a layer with $\Omega(N)$ neurons, $N$ being the number of training samples. Furthermore, there is increasing evidence suggesting that deep networks with sub-linear layer widths are powerful memorizers and optimizers, as long as the number of parameters exceeds the number of samples. Thus, a natural open question is whether the NTK is well conditioned in such a challenging sub-linear setup. In this paper, we answer this question in the affirmative. Our key technical contribution is a lower bound on the smallest NTK eigenvalue for deep networks with the minimum possible over-parameterization: the number of parameters is roughly $\Omega(N)$ and, hence, the number of neurons is as little as $\Omega(\sqrt{N})$. To showcase the applicability of our NTK bounds, we provide two results concerning memorization capacity and optimization guarantees for gradient descent training.

Test-time Batch Normalization

Deep neural networks often suffer the data distribution shift between training and testing, and the batch statistics are observed to reflect the shift. In this paper, targeting of alleviating distribution shift in test time, we revisit the batch normalization (BN) in the training process and reveals two key insights benefiting test-time optimization: $(i)$ preserving the same gradient backpropagation form as training, and $(ii)$ using dataset-level statistics for robust optimization and inference. Based on the two insights, we propose a novel test-time BN layer design, GpreBN, which is optimized during testing by minimizing Entropy loss. We verify the effectiveness of our method on two typical settings with distribution shift, i.e., domain generalization and robustness tasks. Our GpreBN significantly improves the test-time performance and achieves the state of the art results.

How to Guide Adaptive Depth Sampling?

Recent advances in depth sensing technologies allow fast electronic maneuvering of the laser beam, as opposed to fixed mechanical rotations. This will enable future sensors, in principle, to vary in real-time the sampling pattern. We examine here the abstract problem of whether adapting the sampling pattern for a given frame can reduce the reconstruction error or allow a sparser pattern. We propose a constructive generic method to guide adaptive depth sampling algorithms. Given a sampling budget B, a depth predictor P and a desired quality measure M, we propose an Importance Map that highlights important sampling locations. This map is defined for a given frame as the per-pixel expected value of M produced by the predictor P, given a pattern of B random samples. This map can be well estimated in a training phase. We show that a neural network can learn to produce a highly faithful Importance Map, given an RGB image. We then suggest an algorithm to produce a sampling pattern for the scene, which is denser in regions that are harder to reconstruct. The sampling strategy of our modular framework can be adjusted according to hardware limitations, type of depth predictor, and any custom reconstruction error measure that should be minimized. We validate through simulations that our approach outperforms grid and random sampling patterns as well as recent state-of-the-art adaptive algorithms.

The Fairness of Credit Scoring Models

In credit markets, screening algorithms aim to discriminate between good-type and bad-type borrowers. However, when doing so, they also often discriminate between individuals sharing a protected attribute (e.g. gender, age, racial origin) and the rest of the population. In this paper, we show how (1) to test whether there exists a statistically significant difference between protected and unprotected groups, which we call lack of fairness and (2) to identify the variables that cause the lack of fairness. We then use these variables to optimize the fairness-performance trade-off. Our framework provides guidance on how algorithmic fairness can be monitored by lenders, controlled by their regulators, and improved for the benefit of protected groups.

A Proximal Algorithm for Sampling from Non-convex Potentials

We study sampling problems associated with non-convex potentials that meanwhile lack smoothness. In particular, we consider target distributions that satisfy either logarithmic-Sobolev inequality or Poincar\'e inequality. Rather than smooth, the potentials are assumed to be semi-smooth or the summation of multiple semi-smooth functions. We develop a sampling algorithm that resembles proximal algorithms in optimization for this challenging sampling task. Our algorithm is based on a special case of Gibbs sampling known as the alternating sampling framework (ASF). The key contribution of this work is a practical realization of the ASF based on rejection sampling in the non-convex and semi-smooth setting. This work extends the recent algorithm in \cite{LiaChe21,LiaChe22} for non-smooth/semi-smooth log-concave distribution to the setting with non-convex potentials. In almost all the cases of sampling considered in this work, our proximal sampling algorithm achieves better complexity than all existing methods.

Bayesian Active Learning with Fully Bayesian Gaussian Processes

The bias-variance trade-off is a well-known problem in machine learning that only gets more pronounced the less available data there is. In active learning, where labeled data is scarce or difficult to obtain, neglecting this trade-off can cause inefficient and non-optimal querying, leading to unnecessary data labeling. In this paper, we focus on active learning with Gaussian Processes (GPs). For the GP, the bias-variance trade-off is made by optimization of the two hyperparameters: the length scale and noise-term. Considering that the optimal mode of the joint posterior of the hyperparameters is equivalent to the optimal bias-variance trade-off, we approximate this joint posterior and utilize it to design two new acquisition functions. The first one is a Bayesian variant of Query-by-Committee (B-QBC), and the second is an extension that explicitly minimizes the predictive variance through a Query by Mixture of Gaussian Processes (QB-MGP) formulation. Across six common simulators, we empirically show that B-QBC, on average, achieves the best marginal likelihood, whereas QB-MGP achieves the best predictive performance. We show that incorporating the bias-variance trade-off in the acquisition functions mitigates unnecessary and expensive data labeling.

Towards the Generation of Synthetic Images of Palm Vein Patterns: A Review

With the recent success of computer vision and deep learning, remarkable progress has been achieved on automatic personal recognition using vein biometrics. However, collecting large-scale real-world training data for palm vein recognition has turned out to be challenging, mainly due to the noise and irregular variations included at the time of acquisition. Meanwhile, existing palm vein recognition datasets are usually collected under near-infrared light, lacking detailed annotations on attributes (e.g., pose), so the influences of different attributes on vein recognition have been poorly investigated. Therefore, this paper examines the suitability of synthetic vein images generated to compensate for the urgent lack of publicly available large-scale datasets. Firstly, we present an overview of recent research progress on palm vein recognition, from the basic background knowledge to vein anatomical structure, data acquisition, public database, and quality assessment procedures. Then, we focus on the state-of-the-art methods that have allowed the generation of vascular structures for biometric purposes and the modeling of biological networks with their respective application domains. In addition, we review the existing research on the generation of style transfer and biological nature-based synthetic palm vein image algorithms. Afterward, we formalize a general flowchart for the creation of a synthetic database comparing real palm vein images and generated synthetic samples to obtain some understanding into the development of the realistic vein imaging system. Ultimately, we conclude by discussing the challenges, insights, and future perspectives in generating synthetic palm vein images for further works.

Task Relabelling for Multi-task Transfer using Successor Features

Deep Reinforcement Learning has been very successful recently with various works on complex domains. Most works are concerned with learning a single policy that solves the target task, but is fixed in the sense that if the environment changes the agent is unable to adapt to it. Successor Features (SFs) proposes a mechanism that allows learning policies that are not tied to any particular reward function. In this work we investigate how SFs may be pre-trained without observing any reward in a custom environment that features resource collection, traps and crafting. After pre-training we expose the SF agents to various target tasks and see how well they can transfer to new tasks. Transferring is done without any further training on the SF agents, instead just by providing a task vector. For training the SFs we propose a task relabelling method which greatly improves the agent's performance.

Diversity vs. Recognizability: Human-like generalization in one-shot generative models

Robust generalization to new concepts has long remained a distinctive feature of human intelligence. However, recent progress in deep generative models has now led to neural architectures capable of synthesizing novel instances of unknown visual concepts from a single training example. Yet, a more precise comparison between these models and humans is not possible because existing performance metrics for generative models (i.e., FID, IS, likelihood) are not appropriate for the one-shot generation scenario. Here, we propose a new framework to evaluate one-shot generative models along two axes: sample recognizability vs. diversity (i.e., intra-class variability). Using this framework, we perform a systematic evaluation of representative one-shot generative models on the Omniglot handwritten dataset. We first show that GAN-like and VAE-like models fall on opposite ends of the diversity-recognizability space. Extensive analyses of the effect of key model parameters further revealed that spatial attention and context integration have a linear contribution to the diversity-recognizability trade-off. In contrast, disentanglement transports the model along a parabolic curve that could be used to maximize recognizability. Using the diversity-recognizability framework, we were able to identify models and parameters that closely approximate human data.

AutoFedNLP: An efficient FedNLP framework

Transformer-based pre-trained models have revolutionized NLP for superior performance and generality. Fine-tuning pre-trained models for downstream tasks often require private data, for which federated learning is the de-facto approach (i.e., FedNLP). However, our measurements show that FedNLP is prohibitively slow due to the large model sizes and the resultant high network/computation cost. Towards practical FedNLP, we identify as the key building blocks adapters, small bottleneck modules inserted at a variety of model layers. A key challenge is to properly configure the depth and width of adapters, to which the training speed and efficiency is highly sensitive. No silver-bullet configuration exists: the optimal choice varies across downstream NLP tasks, desired model accuracy, and client resources. A silver-bullet configuration does not exist and a non-optimal configuration could significantly slow down the training. To automate adapter configuration, we propose AutoFedNLP, a framework that enhances the existing FedNLP with two novel designs. First, AutoFedNLP progressively upgrades the adapter configuration throughout a training session. Second, AutoFedNLP continuously profiles future adapter configurations by allocating participant devices to trial groups. To minimize client-side computations, AutoFedNLP exploits the fact that a FedNLP client trains on the same samples repeatedly between consecutive changes of adapter configurations, and caches computed activations on clients. Extensive experiments show that AutoFedNLP can reduce FedNLP's model convergence delay to no more than several hours, which is up to 155.5$\times$ faster compared to vanilla FedNLP and 48$\times$ faster compared to strong baselines.

Swapping Semantic Contents for Mixing Images

Deep architecture have proven capable of solving many tasks provided a sufficient amount of labeled data. In fact, the amount of available labeled data has become the principal bottleneck in low label settings such as Semi-Supervised Learning. Mixing Data Augmentations do not typically yield new labeled samples, as indiscriminately mixing contents creates between-class samples. In this work, we introduce the SciMix framework that can learn to generator to embed a semantic style code into image backgrounds, we obtain new mixing scheme for data augmentation. We then demonstrate that SciMix yields novel mixed samples that inherit many characteristics from their non-semantic parents. Afterwards, we verify those samples can be used to improve the performance semi-supervised frameworks like Mean Teacher or Fixmatch, and even fully supervised learning on a small labeled dataset.

Machine Learning for Combinatorial Optimisation of Partially-Specified Problems: Regret Minimisation as a Unifying Lens

It is increasingly common to solve combinatorial optimisation problems that are partially-specified. We survey the case where the objective function or the relations between variables are not known or are only partially specified. The challenge is to learn them from available data, while taking into account a set of hard constraints that a solution must satisfy, and that solving the optimisation problem (esp. during learning) is computationally very demanding. This paper overviews four seemingly unrelated approaches, that can each be viewed as learning the objective function of a hard combinatorial optimisation problem: 1) surrogate-based optimisation, 2) empirical model learning, 3) decision-focused learning (`predict + optimise'), and 4) structured-output prediction. We formalise each learning paradigm, at first in the ways commonly found in the literature, and then bring the formalisations together in a compatible way using regret. We discuss the differences and interactions between these frameworks, highlight the opportunities for cross-fertilization and survey open directions.

Revisiting GANs by Best-Response Constraint: Perspective, Methodology, and Application

In past years, the minimax type single-level optimization formulation and its variations have been widely utilized to address Generative Adversarial Networks (GANs). Unfortunately, it has been proved that these alternating learning strategies cannot exactly reveal the intrinsic relationship between the generator and discriminator, thus easily result in a series of issues, including mode collapse, vanishing gradients and oscillations in the training phase, etc. In this work, by investigating the fundamental mechanism of GANs from the perspective of hierarchical optimization, we propose Best-Response Constraint (BRC), a general learning framework, that can explicitly formulate the potential dependency of the generator on the discriminator. Rather than adopting these existing time-consuming bilevel iterations, we design an implicit gradient scheme with outer-product Hessian approximation as our fast solution strategy. \emph{Noteworthy, we demonstrate that even with different motivations and formulations, a variety of existing GANs ALL can be uniformly improved by our flexible BRC methodology.} Extensive quantitative and qualitative experimental results verify the effectiveness, flexibility and stability of our proposed framework.

The developmental trajectory of object recognition robustness: children are like small adults but unlike big deep neural networks

In laboratory object recognition tasks based on undistorted photographs, both adult humans and Deep Neural Networks (DNNs) perform close to ceiling. Unlike adults', whose object recognition performance is robust against a wide range of image distortions, DNNs trained on standard ImageNet (1.3M images) perform poorly on distorted images. However, the last two years have seen impressive gains in DNN distortion robustness, predominantly achieved through ever-increasing large-scale datasets$\unicode{x2014}$orders of magnitude larger than ImageNet. While this simple brute-force approach is very effective in achieving human-level robustness in DNNs, it raises the question of whether human robustness, too, is simply due to extensive experience with (distorted) visual input during childhood and beyond. Here we investigate this question by comparing the core object recognition performance of 146 children (aged 4$\unicode{x2013}$15) against adults and against DNNs. We find, first, that already 4$\unicode{x2013}$6 year-olds showed remarkable robustness to image distortions and outperform DNNs trained on ImageNet. Second, we estimated the number of $\unicode{x201C}$images$\unicode{x201D}$ children have been exposed to during their lifetime. Compared to various DNNs, children's high robustness requires relatively little data. Third, when recognizing objects children$\unicode{x2014}$like adults but unlike DNNs$\unicode{x2014}$rely heavily on shape but not on texture cues. Together our results suggest that the remarkable robustness to distortions emerges early in the developmental trajectory of human object recognition and is unlikely the result of a mere accumulation of experience with distorted visual input. Even though current DNNs match human performance regarding robustness they seem to rely on different and more data-hungry strategies to do so.

Towards efficient feature sharing in MIMO architectures

Multi-input multi-output architectures propose to train multiple subnetworks within one base network and then average the subnetwork predictions to benefit from ensembling for free. Despite some relative success, these architectures are wasteful in their use of parameters. Indeed, we highlight in this paper that the learned subnetwork fail to share even generic features which limits their applicability on smaller mobile and AR/VR devices. We posit this behavior stems from an ill-posed part of the multi-input multi-output framework. To solve this issue, we propose a novel unmixing step in MIMO architectures that allows subnetworks to properly share features. Preliminary experiments on CIFAR-100 show our adjustments allow feature sharing and improve model performance for small architectures.

FedNoiL: A Simple Two-Level Sampling Method for Federated Learning with Noisy Labels

Federated learning (FL) aims at training a global model on the server side while the training data are collected and located at the local devices. Hence, the labels in practice are usually annotated by clients of varying expertise or criteria and thus contain different amounts of noises. Local training on noisy labels can easily result in overfitting to noisy labels, which is devastating to the global model through aggregation. Although recent robust FL methods take malicious clients into account, they have not addressed local noisy labels on each device and the impact to the global model. In this paper, we develop a simple two-level sampling method "FedNoiL" that (1) selects clients for more robust global aggregation on the server; and (2) selects clean labels and correct pseudo-labels at the client end for more robust local training. The sampling probabilities are built upon clean label detection by the global model. Moreover, we investigate different schedules changing the local epochs between aggregations over the course of FL, which notably improves the communication and computation efficiency in noisy label setting. In experiments with homogeneous/heterogeneous data distributions and noise ratios, we observed that direct combinations of SOTA FL methods with SOTA noisy-label learning methods can easily fail but our method consistently achieves better and robust performance.

LeNSE: Learning To Navigate Subgraph Embeddings for Large-Scale Combinatorial Optimisation

Combinatorial Optimisation problems arise in several application domains and are often formulated in terms of graphs. Many of these problems are NP-hard, but exact solutions are not always needed. Several heuristics have been developed to provide near-optimal solutions; however, they do not typically scale well with the size of the graph. We propose a low-complexity approach for identifying a (possibly much smaller) subgraph of the original graph where the heuristics can be run in reasonable time and with a high likelihood of finding a global near-optimal solution. The core component of our approach is LeNSE, a reinforcement learning algorithm that learns how to navigate the space of possible subgraphs using an Euclidean subgraph embedding as its map. To solve CO problems, LeNSE is provided with a discriminative embedding trained using any existing heuristics using only on a small portion of the original graph. When tested on three problems (vertex cover, max-cut and influence maximisation) using real graphs with up to $10$ million edges, LeNSE identifies small subgraphs yielding solutions comparable to those found by running the heuristics on the entire graph, but at a fraction of the total run time.

Visual Concepts Tokenization

Obtaining the human-like perception ability of abstracting visual concepts from concrete pixels has always been a fundamental and important target in machine learning research fields such as disentangled representation learning and scene decomposition. Towards this goal, we propose an unsupervised transformer-based Visual Concepts Tokenization framework, dubbed VCT, to perceive an image into a set of disentangled visual concept tokens, with each concept token responding to one type of independent visual concept. Particularly, to obtain these concept tokens, we only use cross-attention to extract visual information from the image tokens layer by layer without self-attention between concept tokens, preventing information leakage across concept tokens. We further propose a Concept Disentangling Loss to facilitate that different concept tokens represent independent visual concepts. The cross-attention and disentangling loss play the role of induction and mutual exclusion for the concept tokens, respectively. Extensive experiments on several popular datasets verify the effectiveness of VCT on the tasks of disentangled representation learning and scene decomposition. VCT achieves the state of the art results by a large margin.

Kernel Normalized Convolutional Networks

Existing deep convolutional neural network (CNN) architectures frequently rely upon batch normalization (BatchNorm) to effectively train the model. BatchNorm significantly improves model performance, but performs poorly with smaller batch sizes. To address this limitation, we propose kernel normalization and kernel normalized convolutional layers, and incorporate them into kernel normalized convolutional networks (KNConvNets) as the main building blocks. We implement KNConvNets corresponding to the state-of-the-art CNNs such as ResNet and DenseNet while forgoing BatchNorm layers. Through extensive experiments, we illustrate that KNConvNets consistently outperform their batch, group, and layer normalized counterparts in terms of both accuracy and convergence rate while maintaining competitive computational efficiency.

Semi-self-supervised Automated ICD Coding

Clinical Text Notes (CTNs) contain physicians' reasoning process, written in an unstructured free text format, as they examine and interview patients. In recent years, several studies have been published that provide evidence for the utility of machine learning for predicting doctors' diagnoses from CTNs, a task known as ICD coding. Data annotation is time consuming, particularly when a degree of specialization is needed, as is the case for medical data. This paper presents a method of augmenting a sparsely annotated dataset of Icelandic CTNs with a machine-learned imputation in a semi-self-supervised manner. We train a neural network on a small set of annotated CTNs and use it to extract clinical features from a set of un-annotated CTNs. These clinical features consist of answers to about a thousand potential questions that a physician might find the answers to during a consultation of a patient. The features are then used to train a classifier for the diagnosis of certain types of diseases. We report the results of an evaluation of this data augmentation method over three tiers of data availability to the physician. Our data augmentation method shows a significant positive effect which is diminished when clinical features from the examination of the patient and diagnostics are made available. We recommend our method for augmenting scarce datasets for systems that take decisions based on clinical features that do not include examinations or tests.

A Unified Experiment Design Approach for Cyclic and Acyclic Causal Models

We study experiment design for the unique identification of the causal graph of a system where the graph may contain cycles. The presence of cycles in the structure introduces major challenges for experiment design. Unlike the case of acyclic graphs, learning the skeleton of the causal graph from observational distribution may not be possible. Furthermore, intervening on a variable does not necessarily lead to orienting all the edges incident to it. In this paper, we propose an experiment design approach that can learn both cyclic and acyclic graphs and hence, unifies the task of experiment design for both types of graphs. We provide a lower bound on the number of experiments required to guarantee the unique identification of the causal graph in the worst case, showing that the proposed approach is order-optimal in terms of the number of experiments up to an additive logarithmic term. Moreover, we extend our result to the setting where the size of each experiment is bounded by a constant. For this case, we show that our approach is optimal in terms of the size of the largest experiment required for the unique identification of the causal graph in the worst case.

On Calibration of Ensemble-Based Credal Predictors

In recent years, several classification methods that intend to quantify epistemic uncertainty have been proposed, either by producing predictions in the form of second-order distributions or sets of probability distributions. In this work, we focus on the latter, also called credal predictors, and address the question of how to evaluate them: What does it mean that a credal predictor represents epistemic uncertainty in a faithful manner? To answer this question, we refer to the notion of calibration of probabilistic predictors and extend it to credal predictors. Broadly speaking, we call a credal predictor calibrated if it returns sets that cover the true conditional probability distribution. To verify this property for the important case of ensemble-based credal predictors, we propose a novel nonparametric calibration test that generalizes an existing test for probabilistic predictors to the case of credal predictors. Making use of this test, we empirically show that credal predictors based on deep neural networks are often not well calibrated.

Deployment of Energy-Efficient Deep Learning Models on Cortex-M based Microcontrollers using Deep Compression

Large Deep Neural Networks (DNNs) are the backbone of today's artificial intelligence due to their ability to make accurate predictions when being trained on huge datasets. With advancing technologies, such as the Internet of Things, interpreting large quantities of data generated by sensors is becoming an increasingly important task. However, in many applications not only the predictive performance but also the energy consumption of deep learning models is of major interest. This paper investigates the efficient deployment of deep learning models on resource-constrained microcontroller architectures via network compression. We present a methodology for the systematic exploration of different DNN pruning, quantization, and deployment strategies, targeting different ARM Cortex-M based low-power systems. The exploration allows to analyze trade-offs between key metrics such as accuracy, memory consumption, execution time, and power consumption. We discuss experimental results on three different DNN architectures and show that we can compress them to below 10\% of their original parameter count before their predictive quality decreases. This also allows us to deploy and evaluate them on Cortex-M based microcontrollers.

Unintended memorisation of unique features in neural networks

Neural networks pose a privacy risk due to their propensity to memorise and leak training data. We show that unique features occurring only once in training data are memorised by discriminative multi-layer perceptrons and convolutional neural networks trained on benchmark imaging datasets. We design our method for settings where sensitive training data is not available, for example medical imaging. Our setting knows the unique feature, but not the training data, model weights or the unique feature's label. We develop a score estimating a model's sensitivity to a unique feature by comparing the KL divergences of the model's output distributions given modified out-of-distribution images. We find that typical strategies to prevent overfitting do not prevent unique feature memorisation. And that images containing a unique feature are highly influential, regardless of the influence the images's other features. We also find a significant variation in memorisation with training seed. These results imply that neural networks pose a privacy risk to rarely occurring private information. This risk is more pronounced in healthcare applications since sensitive patient information can be memorised when it remains in training data due to an imperfect data sanitisation process.

On the Prediction Instability of Graph Neural Networks

Instability of trained models, i.e., the dependence of individual node predictions on random factors, can affect reproducibility, reliability, and trust in machine learning systems. In this paper, we systematically assess the prediction instability of node classification with state-of-the-art Graph Neural Networks (GNNs). With our experiments, we establish that multiple instantiations of popular GNN models trained on the same data with the same model hyperparameters result in almost identical aggregated performance but display substantial disagreement in the predictions for individual nodes. We find that up to one third of the incorrectly classified nodes differ across algorithm runs. We identify correlations between hyperparameters, node properties, and the size of the training set with the stability of predictions. In general, maximizing model performance implicitly also reduces model instability.

Understanding and Mitigating the Uncertainty in Zero-Shot Translation

Zero-shot translation is a promising direction for building a comprehensive multilingual neural machine translation (MNMT) system. However, its quality is still not satisfactory due to off-target issues. In this paper, we aim to understand and alleviate the off-target issues from the perspective of uncertainty in zero-shot translation. By carefully examining the translation output and model confidence, we identify two uncertainties that are responsible for the off-target issues, namely, extrinsic data uncertainty and intrinsic model uncertainty. Based on the observations, we propose two light-weight and complementary approaches to denoise the training data for model training, and mask out the vocabulary of the off-target languages in inference. Extensive experiments on both balanced and unbalanced datasets show that our approaches significantly improve the performance of zero-shot translation over strong MNMT baselines. Qualitative analyses provide insights into where our approaches reduce off-target translations

The Unreasonable Effectiveness of Deep Evidential Regression

There is a significant need for principled uncertainty reasoning in machine learning systems as they are increasingly deployed in safety-critical domains. A new approach with uncertainty-aware regression-based neural networks (NNs), based on learning evidential distributions for aleatoric and epistemic uncertainties, shows promise over traditional deterministic methods and typical Bayesian NNs, notably with the capabilities to disentangle aleatoric and epistemic uncertainties. Despite some empirical success of Deep Evidential Regression (DER), there are important gaps in the mathematical foundation that raise the question of why the proposed technique seemingly works. We detail the theoretical shortcomings and analyze the performance on synthetic and real-world data sets, showing that Deep Evidential Regression is a heuristic rather than an exact uncertainty quantification. We go on to propose corrections and redefinitions of how aleatoric and epistemic uncertainties should be extracted from NNs.

Leveraging Relational Information for Learning Weakly Disentangled Representations

Disentanglement is a difficult property to enforce in neural representations. This might be due, in part, to a formalization of the disentanglement problem that focuses too heavily on separating relevant factors of variation of the data in single isolated dimensions of the neural representation. We argue that such a definition might be too restrictive and not necessarily beneficial in terms of downstream tasks. In this work, we present an alternative view over learning (weakly) disentangled representations, which leverages concepts from relational learning. We identify the regions of the latent space that correspond to specific instances of generative factors, and we learn the relationships among these regions in order to perform controlled changes to the latent codes. We also introduce a compound generative model that implements such a weak disentanglement approach. Our experiments shows that the learned representations can separate the relevant factors of variation in the data, while preserving the information needed for effectively generating high quality data samples.

A Case of Exponential Convergence Rates for SVM

Classification is often the first problem described in introductory machine learning classes. Generalization guarantees of classification have historically been offered by Vapnik-Chervonenkis theory. Yet those guarantees are based on intractable algorithms, which has led to the theory of surrogate methods in classification. Guarantees offered by surrogate methods are based on calibration inequalities, which have been shown to be highly sub-optimal under some margin conditions, failing short to capture exponential convergence phenomena. Those "super" fast rates are becoming to be well understood for smooth surrogates, but the picture remains blurry for non-smooth losses such as the hinge loss, associated with the renowned support vector machines. In this paper, we present a simple mechanism to obtain fast convergence rates and we investigate its usage for SVM. In particular, we show that SVM can exhibit exponential convergence rates even without assuming the hard Tsybakov margin condition.

Towards Extremely Fast Bilevel Optimization with Self-governed Convergence Guarantees

Gradient methods have become mainstream techniques for Bi-Level Optimization (BLO) in learning and vision fields. The validity of existing works heavily relies on solving a series of approximation subproblems with extraordinarily high accuracy. Unfortunately, to achieve the approximation accuracy requires executing a large quantity of time-consuming iterations and computational burden is naturally caused. This paper is thus devoted to address this critical computational issue. In particular, we propose a single-level formulation to uniformly understand existing explicit and implicit Gradient-based BLOs (GBLOs). This together with our designed counter-example can clearly illustrate the fundamental numerical and theoretical issues of GBLOs and their naive accelerations. By introducing the dual multipliers as a new variable, we then establish Bilevel Alternating Gradient with Dual Correction (BAGDC), a general framework, which significantly accelerates different categories of existing methods by taking specific settings. A striking feature of our convergence result is that, compared to those original unaccelerated GBLO versions, the fast BAGDC admits a unified non-asymptotic convergence theory towards stationarity. A variety of numerical experiments have also been conducted to demonstrate the superiority of the proposed algorithmic framework.

MaskGAE: Masked Graph Modeling Meets Graph Autoencoders

We present masked graph autoencoder (MaskGAE), a self-supervised learning framework for graph-structured data. Different from previous graph autoencoders (GAEs), MaskGAE adopts masked graph modeling (MGM) as a principled pretext task: masking a portion of edges and attempting to reconstruct the missing part with partially visible, unmasked graph structure. To understand whether MGM can help GAEs learn better representations, we provide both theoretical and empirical evidence to justify the benefits of this pretext task. Theoretically, we establish the connections between GAEs and contrastive learning, showing that MGM significantly improves the self-supervised learning scheme of GAEs. Empirically, we conduct extensive experiments on a number of benchmark datasets, demonstrating the superiority of MaskGAE over several state-of-the-arts on both link prediction and node classification tasks. Our code is publicly available at \url{https://github.com/EdisonLeeeee/MaskGAE}.

The Sufficiency of Off-policyness: PPO is insufficient according to an Off-policy Measure

One of the major difficulties of reinforcement learning is learning from {\em off-policy} samples, which are collected by a different policy (behavior policy) from what the algorithm evaluates (the target policy). Off-policy learning needs to correct the distribution of the samples from the behavior policy towards that of the target policy. Unfortunately, important sampling has an inherent high variance issue which leads to poor gradient estimation in policy gradient methods. We focus on an off-policy Actor-Critic architecture, and propose a novel method, called Preconditioned Proximal Policy Optimization (P3O), which can control the high variance of importance sampling by applying a preconditioner to the Conservative Policy Iteration (CPI) objective. {\em This preconditioning uses the sigmoid function in a special way that when there is no policy change, the gradient is maximal and hence policy gradient will drive a big parameter update for efficient exploration of the parameter space}. This is a novel exploration method that has not been studied before given that existing exploration methods are based on the novelty of states and actions. We compare with several best-performing algorithms on both discrete and continuous tasks and the results showed that {\em PPO is insufficient in off-policyness}, and our P3O is {\em more off-policy} than PPO according to the "off-policyness" measured by the DEON metric, and P3O explores in a larger policy space than PPO. Results also show that our P3O maximizes the CPI objective better than PPO during the training process.

ExMo: Explainable AI Model using Inverse Frequency Decision Rules

In this paper, we present a novel method to compute decision rules to build a more accurate interpretable machine learning model, denoted as ExMo. The ExMo interpretable machine learning model consists of a list of IF...THEN... statements with a decision rule in the condition. This way, ExMo naturally provides an explanation for a prediction using the decision rule that was triggered. ExMo uses a new approach to extract decision rules from the training data using term frequency-inverse document frequency (TF-IDF) features. With TF-IDF, decision rules with feature values that are more relevant to each class are extracted. Hence, the decision rules obtained by ExMo can distinguish the positive and negative classes better than the decision rules used in the existing Bayesian Rule List (BRL) algorithm, obtained using the frequent pattern mining approach. The paper also shows that ExMo learns a qualitatively better model than BRL. Furthermore, ExMo demonstrates that the textual explanation can be provided in a human-friendly way so that the explanation can be easily understood by non-expert users. We validate ExMo on several datasets with different sizes to evaluate its efficacy. Experimental validation on a real-world fraud detection application shows that ExMo is 20% more accurate than BRL and that it achieves accuracy similar to those of deep learning models.

Towards biologically plausible Dreaming and Planning

Humans and animals can learn new skills after practicing for a few hours, while current reinforcement learning algorithms require a large amount of data to achieve good performances. Recent model-based approaches show promising results by reducing the number of necessary interactions with the environment to learn a desirable policy. However, these methods require biological implausible ingredients, such as the detailed storage of older experiences, and long periods of offline learning. The optimal way to learn and exploit word-models is still an open question. Taking inspiration from biology, we suggest that dreaming might be an efficient expedient to use an inner model. We propose a two-module (agent and model) neural network in which "dreaming" (living new experiences in a model-based simulated environment) significantly boosts learning. We also explore "planning", an online alternative to dreaming, that shows comparable performances. Importantly, our model does not require the detailed storage of experiences, and learns online the world-model. This is a key ingredient for biological plausibility and implementability (e.g., in neuromorphic hardware). Our network is composed of spiking neurons, further increasing the energetic efficiency and the plausibility of the model. To our knowledge, there are no previous works proposing biologically plausible model-based reinforcement learning in recurrent spiking networks. Our work is a step toward building efficient neuromorphic systems for autonomous robots, capable of learning new skills in real-world environments. Even when the environment is no longer accessible, the robot optimizes learning by "reasoning" in its own "mind". These approaches are of great relevance when the acquisition from the environment is slow, expensive (robotics) or unsafe (autonomous driving).

Posterior Refinement Improves Sample Efficiency in Bayesian Neural Networks

Monte Carlo (MC) integration is the de facto method for approximating the predictive distribution of Bayesian neural networks (BNNs). But, even with many MC samples, Gaussian-based BNNs could still yield bad predictive performance due to the posterior approximation's error. Meanwhile, alternatives to MC integration tend to be more expensive or biased. In this work, we experimentally show that the key to good MC-approximated predictive distributions is the quality of the approximate posterior itself. However, previous methods for obtaining accurate posterior approximations are expensive and non-trivial to implement. We, therefore, propose to refine Gaussian approximate posteriors with normalizing flows. When applied to last-layer BNNs, it yields a simple \emph{post hoc} method for improving pre-existing parametric approximations. We show that the resulting posterior approximation is competitive with even the gold-standard full-batch Hamiltonian Monte Carlo.

Exploring Extreme Parameter Compression for Pre-trained Language Models

Recent work explored the potential of large-scale Transformer-based pre-trained models, especially Pre-trained Language Models (PLMs) in natural language processing. This raises many concerns from various perspectives, e.g., financial costs and carbon emissions. Compressing PLMs like BERT with negligible performance loss for faster inference and cheaper deployment has attracted much attention. In this work, we aim to explore larger compression ratios for PLMs, among which tensor decomposition is a potential but under-investigated one. Two decomposition and reconstruction protocols are further proposed to improve the effectiveness and efficiency during compression. Our compressed BERT with ${1}/{7}$ parameters in Transformer layers performs on-par with, sometimes slightly better than the original BERT in GLUE benchmark. A tiny version achieves $96.7\%$ performance of BERT-base with $ {1}/{48} $ encoder parameters (i.e., less than 2M parameters excluding the embedding layer) and $2.7 \times$ faster on inference. To show that the proposed method is orthogonal to existing compression methods like knowledge distillation, we also explore the benefit of the proposed method on a distilled BERT.

Survey on Fair Reinforcement Learning: Theory and Practice

Fairness-aware learning aims at satisfying various fairness constraints in addition to the usual performance criteria via data-driven machine learning techniques. Most of the research in fairness-aware learning employs the setting of fair-supervised learning. However, many dynamic real-world applications can be better modeled using sequential decision-making problems and fair reinforcement learning provides a more suitable alternative for addressing these problems. In this article, we provide an extensive overview of fairness approaches that have been implemented via a reinforcement learning (RL) framework. We discuss various practical applications in which RL methods have been applied to achieve a fair solution with high accuracy. We further include various facets of the theory of fair reinforcement learning, organizing them into single-agent RL, multi-agent RL, long-term fairness via RL, and offline learning. Moreover, we highlight a few major issues to explore in order to advance the field of fair-RL, namely - i) correcting societal biases, ii) feasibility of group fairness or individual fairness, and iii) explainability in RL. Our work is beneficial for both researchers and practitioners as we discuss articles providing mathematical guarantees as well as articles with empirical studies on real-world problems.

Trend analysis and forecasting air pollution in Rwanda

Air pollution is a major public health problem worldwide although the lack of data is a global issue for most low and middle income countries. Ambient air pollution in the form of fine particulate matter (PM2.5) exceeds the World Health Organization guidelines in Rwanda with a daily average of around 42.6 microgram per meter cube. Monitoring and mitigation strategies require an expensive investment in equipment to collect pollution data. Low-cost sensor technology and machine learning methods have appeared as an alternative solution to get reliable information for decision making. This paper analyzes the trend of air pollution in Rwanda and proposes forecasting models suitable to data collected by a network of low-cost sensors deployed in Rwanda.

Towards Consistency in Adversarial Classification

In this paper, we study the problem of consistency in the context of adversarial examples. Specifically, we tackle the following question: can surrogate losses still be used as a proxy for minimizing the $0/1$ loss in the presence of an adversary that alters the inputs at test-time? Different from the standard classification task, this question cannot be reduced to a point-wise minimization problem, and calibration needs not to be sufficient to ensure consistency. In this paper, we expose some pathological behaviors specific to the adversarial problem, and show that no convex surrogate loss can be consistent or calibrated in this context. It is therefore necessary to design another class of surrogate functions that can be used to solve the adversarial consistency issue. As a first step towards designing such a class, we identify sufficient and necessary conditions for a surrogate loss to be calibrated in both the adversarial and standard settings. Finally, we give some directions for building a class of losses that could be consistent in the adversarial framework.

Predicting electrode array impedance after one month from cochlear implantation surgery

Sensorineural hearing loss can be treated using Cochlear implantation. After this surgery using the electrode array impedance measurements, we can check the stability of the impedance value and the dynamic range. Deterioration of speech recognition scores could happen because of increased impedance values. Medicines used to do these measures many times during a year after the surgery. Predicting the electrode impedance could help in taking decisions to help the patient get better hearing. In this research we used a dataset of 80 patients of children who did cochlear implantation using MED-EL FLEX28 electrode array of 12 channels. We predicted the electrode impedance on each channel after 1 month from the surgery date. We used different machine learning algorithms like neural networks and decision trees. Our results indicates that the electrode impedance can be predicted, and the best algorithm is different based on the electrode channel. Also, the accuracy level varies between 66% and 100% based on the electrode channel when accepting an error range between 0 and 3 KO. Further research is required to predict the electrode impedance after three months, six months and one year.

Neural Additive Models for Nowcasting

Deep neural networks (DNNs) are one of the most highlighted methods in machine learning. However, as DNNs are black-box models, they lack explanatory power for their predictions. Recently, neural additive models (NAMs) have been proposed to provide this power while maintaining high prediction performance. In this paper, we propose a novel NAM approach for multivariate nowcasting (NC) problems, which comprise an important focus area of machine learning. For the multivariate time-series data used in NC problems, explanations should be considered for every input value to the variables at distinguishable time steps. By employing generalized additive models, the proposed NAM-NC successfully explains each input value's importance for multiple variables and time steps. Experimental results involving a toy example and two real-world datasets show that the NAM-NC predicts multivariate time-series data as accurately as state-of-the-art neural networks, while also providing the explanatory importance of each input value. We also examine parameter-sharing networks using NAM-NC to decrease their complexity, and NAM-MC's hard-tied feature net extracted explanations with good performance.

Translating Hanja historical documents to understandable Korean and English

The Annals of Joseon Dynasty (AJD) contain the daily records of the Kings of Joseon, the 500-year kingdom preceding the modern nation of Korea. The Annals were originally written in an archaic Korean writing system, `Hanja', and translated into Korean from 1968 to 1993. However, this translation was literal and contained many archaic Korean words; thus, a new expert translation effort began in 2012, completing the records of only one king in a decade. Also, expert translators are working on an English translation, of which only one king's records are available because of the high cost and slow progress. Thus, we propose H2KE, the neural machine translation model that translates Hanja historical documents to understandable Korean and English. Based on the multilingual neural machine translation approach, it translates the historical document written in Hanja, using both the full dataset of outdated Korean translation and a small dataset of recently translated Korean and English. We compare our method with two baselines: one is a recent model that simultaneously learns to restore and translate Hanja historical document and the other is the transformer that trained on newly translated corpora only. The results show that our method significantly outperforms the baselines in terms of BLEU score in both modern Korean and English translations. We also conduct a human evaluation that shows that our translation is preferred over the original expert translation.

Self-Paced Multi-Agent Reinforcement Learning

Curriculum reinforcement learning (CRL) aims to speed up learning of a task by changing gradually the difficulty of the task from easy to hard through control of factors such as initial state or environment dynamics. While automating CRL is well studied in the single-agent setting, in multi-agent reinforcement learning (MARL) an open question is whether control of the number of agents with other factors in a principled manner is beneficial, prior approaches typically relying on hand-crafted heuristics. In addition, how the tasks evolve as the number of agents changes remains understudied, which is critical for scaling to more challenging tasks. We introduce self-paced MARL (SPMARL) that enables optimizing the number of agents with other environment factors in a principled way, and, show that usual assumptions such as that fewer agents make the task always easier are not generally valid. The curriculum induced by SPMARL reveals the evolution of tasks w.r.t. number of agents and experiments show that SPMARL improves the performance when the number of agents sufficiently influences task difficulty.

A Survey of Trustworthy Graph Learning: Reliability, Explainability, and Privacy Protection

Deep graph learning has achieved remarkable progresses in both business and scientific areas ranging from finance and e-commerce, to drug and advanced material discovery. Despite these progresses, how to ensure various deep graph learning algorithms behave in a socially responsible manner and meet regulatory compliance requirements becomes an emerging problem, especially in risk-sensitive domains. Trustworthy graph learning (TwGL) aims to solve the above problems from a technical viewpoint. In contrast to conventional graph learning research which mainly cares about model performance, TwGL considers various reliability and safety aspects of the graph learning framework including but not limited to robustness, explainability, and privacy. In this survey, we provide a comprehensive review of recent leading approaches in the TwGL field from three dimensions, namely, reliability, explainability, and privacy protection. We give a general categorization for existing work and review typical work for each category. To give further insights for TwGL research, we provide a unified view to inspect previous works and build the connection between them. We also point out some important open problems remaining to be solved in the future developments of TwGL.

Constructive Interpretability with CoLabel: Corroborative Integration, Complementary Features, and Collaborative Learning

Machine learning models with explainable predictions are increasingly sought after, especially for real-world, mission-critical applications that require bias detection and risk mitigation. Inherent interpretability, where a model is designed from the ground-up for interpretability, provides intuitive insights and transparent explanations on model prediction and performance. In this paper, we present CoLabel, an approach to build interpretable models with explanations rooted in the ground truth. We demonstrate CoLabel in a vehicle feature extraction application in the context of vehicle make-model recognition (VMMR). CoLabel performs VMMR with a composite of interpretable features such as vehicle color, type, and make, all based on interpretable annotations of the ground truth labels. First, CoLabel performs corroborative integration to join multiple datasets that each have a subset of desired annotations of color, type, and make. Then, CoLabel uses decomposable branches to extract complementary features corresponding to desired annotations. Finally, CoLabel fuses them together for final predictions. During feature fusion, CoLabel harmonizes complementary branches so that VMMR features are compatible with each other and can be projected to the same semantic space for classification. With inherent interpretability, CoLabel achieves superior performance to the state-of-the-art black-box models, with accuracy of 0.98, 0.95, and 0.94 on CompCars, Cars196, and BoxCars116K, respectively. CoLabel provides intuitive explanations due to constructive interpretability, and subsequently achieves high accuracy and usability in mission-critical situations.

The price of ignorance: how much does it cost to forget noise structure in low-rank matrix estimation?

We consider the problem of estimating a rank-1 signal corrupted by structured rotationally invariant noise, and address the following question: how well do inference algorithms perform when the noise statistics is unknown and hence Gaussian noise is assumed? While the matched Bayes-optimal setting with unstructured noise is well understood, the analysis of this mismatched problem is only at its premises. In this paper, we make a step towards understanding the effect of the strong source of mismatch which is the noise statistics. Our main technical contribution is the rigorous analysis of a Bayes estimator and of an approximate message passing (AMP) algorithm, both of which incorrectly assume a Gaussian setup. The first result exploits the theory of spherical integrals and of low-rank matrix perturbations; the idea behind the second one is to design and analyze an artificial AMP which, by taking advantage of the flexibility in the denoisers, is able to "correct" the mismatch. Armed with these sharp asymptotic characterizations, we unveil a rich and often unexpected phenomenology. For example, despite AMP is in principle designed to efficiently compute the Bayes estimator, the former is outperformed by the latter in terms of mean-square error. We show that this performance gap is due to an incorrect estimation of the signal norm. In fact, when the SNR is large enough, the overlaps of the AMP and the Bayes estimator coincide, and they even match those of optimal estimators taking into account the structure of the noise.

Self-Supervised Depth Estimation with Isometric-Self-Sample-Based Learning

Managing the dynamic regions in the photometric loss formulation has been a main issue for handling the self-supervised depth estimation problem. Most previous methods have alleviated this issue by removing the dynamic regions in the photometric loss formulation based on the masks estimated from another module, making it difficult to fully utilize the training images. In this paper, to handle this problem, we propose an isometric self-sample-based learning (ISSL) method to fully utilize the training images in a simple yet effective way. The proposed method provides additional supervision during training using self-generated images that comply with pure static scene assumption. Specifically, the isometric self-sample generator synthesizes self-samples for each training image by applying random rigid transformations on the estimated depth. Thus both the generated self-samples and the corresponding training image always follow the static scene assumption. We show that plugging our ISSL module into several existing models consistently improves the performance by a large margin. In addition, it also boosts the depth accuracy over different types of scene, i.e., outdoor scenes (KITTI and Make3D) and indoor scene (NYUv2), validating its high effectiveness.

RiskLoc: Localization of Multi-dimensional Root Causes by Weighted Risk

Failures and anomalies in large-scale software systems are unavoidable incidents. When an issue is detected, operators need to quickly and correctly identify its location to facilitate a swift repair. In this work, we consider the problem of identifying the root cause set that best explains an anomaly in multi-dimensional time series with categorical attributes. The huge search space is the main challenge, even for a small number of attributes and small value sets, the number of theoretical combinations is too large to brute force. Previous approaches have thus focused on reducing the search space, but they all suffer from various issues, requiring extensive manual parameter tuning, being too slow and thus impractical, or being incapable of finding more complex root causes. We propose RiskLoc to solve the problem of multidimensional root cause localization. RiskLoc applies a 2-way partitioning scheme and assigns element weights that linearly increase with the distance from the partitioning point. A risk score is assigned to each element that integrates two factors, 1) its weighted proportion within the abnormal partition, and 2) the relative change in the deviation score adjusted for the ripple effect property. Extensive experiments on multiple datasets verify the effectiveness and efficiency of RiskLoc, and for a comprehensive evaluation, we introduce three synthetically generated datasets that complement existing datasets. We demonstrate that RiskLoc consistently outperforms state-of-the-art baselines, especially in more challenging root cause scenarios, with gains in F1-score up to 57% over the second-best approach with comparable running times.

MPI: Evaluating and Inducing Personality in Pre-trained Language Models

Originated as a philosophical quest, personality discerns how individuals differ from each other in terms of thinking, feeling, and behaving. Towards building social machines that work with humans on a daily basis, we are motivated to ask: (1) Do existing pre-trained language models possess personality, akin to their human counterpart? If so, (2) how can we evaluate them? Further, given this evaluation framework, (3) how can we induce a certain personality in a fully controllable fashion? To tackle these three questions, we propose the Machine Personality Inventory (MPI) dataset for evaluating the machine personality; MPI follows standardized personality tests, built upon the Big Five Personality Factors (Big Five) theory and personality assessment inventories. By evaluating models with MPI, we provide the first piece of evidence showing the existence of personality in pre-trained language models. We further devise a Chain Prompting method to induce the language model with a specific personality in a controllable manner, capable of producing diversified behaviors. We hope to shed light on future studies by adopting personality as the essential psychological guidance for various downstream tasks, building more human-like and in situ dialogue agents.

Planning with Diffusion for Flexible Behavior Synthesis

Model-based reinforcement learning methods often use learning only for the purpose of estimating an approximate dynamics model, offloading the rest of the decision-making work to classical trajectory optimizers. While conceptually simple, this combination has a number of empirical shortcomings, suggesting that learned models may not be well-suited to standard trajectory optimization. In this paper, we consider what it would look like to fold as much of the trajectory optimization pipeline as possible into the modeling problem, such that sampling from the model and planning with it become nearly identical. The core of our technical approach lies in a diffusion probabilistic model that plans by iteratively denoising trajectories. We show how classifier-guided sampling and image inpainting can be reinterpreted as coherent planning strategies, explore the unusual and useful properties of diffusion-based planning methods, and demonstrate the effectiveness of our framework in control settings that emphasize long-horizon decision-making and test-time flexibility.

Set-based Meta-Interpolation for Few-Task Meta-Learning

Meta-learning approaches enable machine learning systems to adapt to new tasks given few examples by leveraging knowledge from related tasks. However, a large number of meta-training tasks are still required for generalization to unseen tasks during meta-testing, which introduces a critical bottleneck for real-world problems that come with only few tasks, due to various reasons including the difficulty and cost of constructing tasks. Recently, several task augmentation methods have been proposed to tackle this issue using domain-specific knowledge to design augmentation techniques to densify the meta-training task distribution. However, such reliance on domain-specific knowledge renders these methods inapplicable to other domains. While Manifold Mixup based task augmentation methods are domain-agnostic, we empirically find them ineffective on non-image domains. To tackle these limitations, we propose a novel domain-agnostic task augmentation method, Meta-Interpolation, which utilizes expressive neural set functions to densify the meta-training task distribution using bilevel optimization. We empirically validate the efficacy of Meta-Interpolation on eight datasets spanning across various domains such as image classification, molecule property prediction, text classification and speech recognition. Experimentally, we show that Meta-Interpolation consistently outperforms all the relevant baselines. Theoretically, we prove that task interpolation with the set function regularizes the meta-learner to improve generalization.

Nonlinear motion separation via untrained generator networks with disentangled latent space variables and applications to cardiac MRI

In this paper, a nonlinear approach to separate different motion types in video data is proposed. This is particularly relevant in dynamic medical imaging (e.g. PET, MRI), where patient motion poses a significant challenge due to its effects on the image reconstruction as well as for its subsequent interpretation. Here, a new method is proposed where dynamic images are represented as the forward mapping of a sequence of latent variables via a generator neural network. The latent variables are structured so that temporal variations in the data are represented via dynamic latent variables, which are independent of static latent variables characterizing the general structure of the frames. In particular, different kinds of motion are also characterized independently of each other via latent space disentanglement using one-dimensional prior information on all but one of the motion types. This representation allows to freeze any selection of motion types, and to obtain accurate independent representations of other dynamics of interest. Moreover, the proposed algorithm is training-free, i.e., all the network parameters are learned directly from a single video. We illustrate the performance of this method on phantom and real-data MRI examples, where we successfully separate respiratory and cardiac motion.

SafeNet: Mitigating Data Poisoning Attacks on Private Machine Learning

Secure multiparty computation (MPC) has been proposed to allow multiple mutually distrustful data owners to jointly train machine learning (ML) models on their combined data. However, the datasets used for training ML models might be under the control of an adversary mounting a data poisoning attack, and MPC prevents inspecting training sets to detect poisoning. We show that multiple MPC frameworks for private ML training are susceptible to backdoor and targeted poisoning attacks. To mitigate this, we propose SafeNet, a framework for building ensemble models in MPC with formal guarantees of robustness to data poisoning attacks. We extend the security definition of private ML training to account for poisoning and prove that our SafeNet design satisfies the definition. We demonstrate SafeNet's efficiency, accuracy, and resilience to poisoning on several machine learning datasets and models. For instance, SafeNet reduces backdoor attack success from 100% to 0% for a neural network model, while achieving 39x faster training and 36x less communication than the four-party MPC framework of Dalskov et al.

HeadText: Exploring Hands-free Text Entry using Head Gestures by Motion Sensing on a Smart Earpiece

We present HeadText, a hands-free technique on a smart earpiece for text entry by motion sensing. Users input text utilizing only 7 head gestures for key selection, word selection, word commitment and word cancelling tasks. Head gesture recognition is supported by motion sensing on a smart earpiece to capture head moving signals and machine learning algorithms (K-Nearest-Neighbor (KNN) with a Dynamic Time Warping (DTW) distance measurement). A 10-participant user study proved that HeadText could recognize 7 head gestures at an accuracy of 94.29%. After that, the second user study presented that HeadText could achieve a maximum accuracy of 10.65 WPM and an average accuracy of 9.84 WPM for text entry. Finally, we demonstrate potential applications of HeadText in hands-free scenarios for (a). text entry of people with motor impairments, (b). private text entry, and (c). socially acceptable text entry.

FairNorm: Fair and Fast Graph Neural Network Training

Graph neural networks (GNNs) have been demonstrated to achieve state-of-the-art for a number of graph-based learning tasks, which leads to a rise in their employment in various domains. However, it has been shown that GNNs may inherit and even amplify bias within training data, which leads to unfair results towards certain sensitive groups. Meanwhile, training of GNNs introduces additional challenges, such as slow convergence and possible instability. Faced with these limitations, this work proposes FairNorm, a unified normalization framework that reduces the bias in GNN-based learning while also providing provably faster convergence. Specifically, FairNorm employs fairness-aware normalization operators over different sensitive groups with learnable parameters to reduce the bias in GNNs. The design of FairNorm is built upon analyses that illuminate the sources of bias in graph-based learning. Experiments on node classification over real-world networks demonstrate the efficiency of the proposed scheme in improving fairness in terms of statistical parity and equal opportunity compared to fairness-aware baselines. In addition, it is empirically shown that the proposed framework leads to faster convergence compared to the naive baseline where no normalization is employed.

Neuro-Symbolic Regex Synthesis Framework via Neural Example Splitting

Due to the practical importance of regular expressions (regexes, for short), there has been a lot of research to automatically generate regexes from positive and negative string examples. We tackle the problem of learning regexes faster from positive and negative strings by relying on a novel approach called `neural example splitting'. Our approach essentially split up each example string into multiple parts using a neural network trained to group similar substrings from positive strings. This helps to learn a regex faster and, thus, more accurately since we now learn from several short-length strings. We propose an effective regex synthesis framework called `SplitRegex' that synthesizes subregexes from `split' positive substrings and produces the final regex by concatenating the synthesized subregexes. For the negative sample, we exploit pre-generated subregexes during the subregex synthesis process and perform the matching against negative strings. Then the final regex becomes consistent with all negative strings. SplitRegex is a divided-and-conquer framework for learning target regexes; split (=divide) positive strings and infer partial regexes for multiple parts, which is much more accurate than the whole string inferring, and concatenate (=conquer) inferred regexes while satisfying negative strings. We empirically demonstrate that the proposed SplitRegex framework substantially improves the previous regex synthesis approaches over four benchmark datasets.

Diagnosis and Prognosis of COVID-19 Disease Using Routine Blood Values and LogNNet Neural Network

Since February 2020, the world has been engaged in an intense struggle with the COVID-19 dis-ease, and health systems have come under tragic pressure as the disease turned into a pandemic. The aim of this study is to obtain the most effective routine blood values (RBV) in the diagnosis and prognosis of COVID-19 using a backward feature elimination algorithm for the LogNNet reservoir neural network. The first dataset in the study consists of a total of 5296 patients with the same number of negative and positive COVID-19 tests. The LogNNet-model achieved the accuracy rate of 99.5% in the diagnosis of the disease with 46 features and the accuracy of 99.17% with only mean corpuscular hemoglobin concentration, mean corpuscular hemoglobin, and activated partial prothrombin time. The second dataset consists of a total of 3899 patients with a diagnosis of COVID-19 who were treated in hospital, of which 203 were severe patients and 3696 were mild patients. The model reached the accuracy rate of 94.4% in determining the prognosis of the disease with 48 features and the accuracy of 82.7% with only erythrocyte sedimentation rate, neutrophil count, and C reactive protein features. Our method will reduce the negative pressures on the health sector and help doctors to understand the pathogenesis of COVID-19 using the key features. The method is promising to create mobile health monitoring systems in the Internet of Things.

On Tackling Explanation Redundancy in Decision Trees

Decision trees (DTs) epitomize the ideal of interpretability of machine learning (ML) models. The interpretability of decision trees motivates explainability approaches by so-called intrinsic interpretability, and it is at the core of recent proposals for applying interpretable ML models in high-risk applications. The belief in DT interpretability is justified by the fact that explanations for DT predictions are generally expected to be succinct. Indeed, in the case of DTs, explanations correspond to DT paths. Since decision trees are ideally shallow, and so paths contain far fewer features than the total number of features, explanations in DTs are expected to be succinct, and hence interpretable. This paper offers both theoretical and experimental arguments demonstrating that, as long as interpretability of decision trees equates with succinctness of explanations, then decision trees ought not be deemed interpretable. The paper introduces logically rigorous path explanations and path explanation redundancy, and proves that there exist functions for which decision trees must exhibit paths with arbitrarily large explanation redundancy. The paper also proves that only a very restricted class of functions can be represented with DTs that exhibit no explanation redundancy. In addition, the paper includes experimental results substantiating that path explanation redundancy is observed ubiquitously in decision trees, including those obtained using different tree learning algorithms, but also in a wide range of publicly available decision trees. The paper also proposes polynomial-time algorithms for eliminating path explanation redundancy, which in practice require negligible time to compute. Thus, these algorithms serve to indirectly attain irreducible, and so succinct, explanations for decision trees.

Actively Tracking the Optimal Arm in Non-Stationary Environments with Mandatory Probing

We study a novel multi-armed bandit (MAB) setting which mandates the agent to probe all the arms periodically in a non-stationary environment. In particular, we develop \texttt{TS-GE} that balances the regret guarantees of classical Thompson sampling (TS) with the broadcast probing (BP) of all the arms simultaneously in order to actively detect a change in the reward distributions. Once a system-level change is detected, the changed arm is identified by an optional subroutine called group exploration (GE) which scales as $\log_2(K)$ for a $K-$armed bandit setting. We characterize the probability of missed detection and the probability of false-alarm in terms of the environment parameters. The latency of change-detection is upper bounded by $\sqrt{T}$ while within a period of $\sqrt{T}$, all the arms are probed at least once. We highlight the conditions in which the regret guarantee of \texttt{TS-GE} outperforms that of the state-of-the-art algorithms, in particular, \texttt{ADSWITCH} and \texttt{M-UCB}. Furthermore, unlike the existing bandit algorithms, \texttt{TS-GE} can be deployed for applications such as timely status updates, critical control, and wireless energy transfer, which are essential features of next-generation wireless communication networks. We demonstrate the efficacy of \texttt{TS-GE} by employing it in a n industrial internet-of-things (IIoT) network designed for simultaneous wireless information and power transfer (SWIPT).

A General Framework for quantifying Aleatoric and Epistemic uncertainty in Graph Neural Networks

Graph Neural Networks (GNN) provide a powerful framework that elegantly integrates Graph theory with Machine learning for modeling and analysis of networked data. We consider the problem of quantifying the uncertainty in predictions of GNN stemming from modeling errors and measurement uncertainty. We consider aleatoric uncertainty in the form of probabilistic links and noise in feature vector of nodes, while epistemic uncertainty is incorporated via a probability distribution over the model parameters. We propose a unified approach to treat both sources of uncertainty in a Bayesian framework, where Assumed Density Filtering is used to quantify aleatoric uncertainty and Monte Carlo dropout captures uncertainty in model parameters. Finally, the two sources of uncertainty are aggregated to estimate the total uncertainty in predictions of a GNN. Results in the real-world datasets demonstrate that the Bayesian model performs at par with a frequentist model and provides additional information about predictions uncertainty that are sensitive to uncertainties in the data and model.

A Fully Controllable Agent in the Path Planning using Goal-Conditioned Reinforcement Learning

The aim of path planning is to reach the goal from starting point by searching for the route of an agent. In the path planning, the routes may vary depending on the number of variables such that it is important for the agent to reach various goals. Numerous studies, however, have dealt with a single goal that is predefined by the user. In the present study, I propose a novel reinforcement learning framework for a fully controllable agent in the path planning. To do this, I propose a bi-directional memory editing to obtain various bi-directional trajectories of the agent, in which the behavior of the agent and sub-goals are trained on the goal-conditioned RL. As for moving the agent in various directions, I utilize the sub-goals dedicated network, separated from a policy network. Lastly, I present the reward shaping to shorten the number of steps for the agent to reach the goal. In the experimental result, the agent was able to reach the various goals that have never been visited by the agent in the training. We confirmed that the agent could perform difficult missions such as a round trip and the agent used the shorter route with the reward shaping.

Sample Complexity of Learning Heuristic Functions for Greedy-Best-First and A* Search

Greedy best-first search (GBFS) and A* search (A*) are popular algorithms for path-finding on large graphs. Both use so-called heuristic functions, which estimate how close a vertex is to the goal. While heuristic functions have been handcrafted using domain knowledge, recent studies demonstrate that learning heuristic functions from data is effective in many applications. Motivated by this emerging approach, we study the sample complexity of learning heuristic functions for GBFS and A*. We build on a recent framework called \textit{data-driven algorithm design} and evaluate the \textit{pseudo-dimension} of a class of utility functions that measure the performance of parameterized algorithms. Assuming that a vertex set of size $n$ is fixed, we present $\mathrm{O}(n\lg n)$ and $\mathrm{O}(n^2\lg n)$ upper bounds on the pseudo-dimensions for GBFS and A*, respectively, parameterized by heuristic function values. The upper bound for A* can be improved to $\mathrm{O}(n^2\lg d)$ if every vertex has a degree of at most $d$ and to $\mathrm{O}(n \lg n)$ if edge weights are integers bounded by $\mathrm{poly}(n)$. We also give $\Omega(n)$ lower bounds for GBFS and A*, which imply that our bounds for GBFS and A* under the integer-weight condition are tight up to a $\lg n$ factor. Finally, we discuss a case where the performance of A* is measured by the suboptimality and show that we can sometimes obtain a better guarantee by combining a parameter-dependent worst-case bound with a sample complexity bound.

Discrete-Convex-Analysis-Based Framework for Warm-Starting Algorithms with Predictions

Augmenting algorithms with learned predictions is a promising approach for going beyond worst-case bounds. Dinitz, Im, Lavastida, Moseley, and Vassilvitskii~(2021) have demonstrated that a warm start with learned dual solutions can improve the time complexity of the Hungarian method for weighted perfect bipartite matching. We extend and improve their framework in a principled manner via \textit{discrete convex analysis} (DCA), a discrete analog of convex analysis. We show the usefulness of our DCA-based framework by applying it to weighted perfect bipartite matching, weighted matroid intersection, and discrete energy minimization for computer vision. Our DCA-based framework yields time complexity bounds that depend on the $\ell_\infty$-distance from a predicted solution to an optimal solution, which has two advantages relative to the previous $\ell_1$-distance-dependent bounds: time complexity bounds are smaller, and learning of predictions is more sample efficient. We also discuss whether to learn primal or dual solutions from the DCA perspective.

A Correlation Information-based Spatiotemporal Network for Traffic Flow Forecasting

With the growth of transport modes, high traffic forecasting precision is required in intelligent transportation systems. Most previous works utilize the transformer architecture based on graph neural networks and attention mechanisms to discover spatiotemporal dependencies and dynamic relationships. The correlation information among spatiotemporal sequences, however, has not been thoroughly considered. In this paper, we present two elaborate spatiotemporal representations, spatial correlation information (SCorr) and temporal correlation information (TCorr), among spatiotemporal sequences based on the maximal information coefficient. Using SCorr, we propose a novel correlation information-based spatiotemporal network (CorrSTN), including a dynamic graph neural network component incorporating correlation information into the spatial structure effectively and a multi-head attention component utilizing spatial correlation information to extract dynamic temporal dependencies accurately. Using TCorr, we further explore the correlation pattern among different periodic data and then propose a novel data selection scheme to identify the most relevant data. The experimental results on the highway traffic flow (PEMS07 and PEMS08) and metro crowd flow (HZME inflow and outflow) datasets demonstrate that CorrSTN outperforms the state-of-the-art methods in terms of predictive performance. In particular, on the HZME (outflow) dataset, our model makes significant improvements compared with the latest model ASTGNN by 12.7%, 14.4% and 27.4% in the metrics of MAE, RMSE and MAPE, respectively.

Learning to Reverse DNNs from AI Programs Automatically

With the privatization deployment of DNNs on edge devices, the security of on-device DNNs has raised significant concern. To quantify the model leakage risk of on-device DNNs automatically, we propose NNReverse, the first learning-based method which can reverse DNNs from AI programs without domain knowledge. NNReverse trains a representation model to represent the semantics of binary code for DNN layers. By searching the most similar function in our database, NNReverse infers the layer type of a given function's binary code. To represent assembly instructions semantics precisely, NNReverse proposes a more fine-grained embedding model to represent the textual and structural-semantic of assembly functions.

Explainable Supervised Domain Adaptation

Domain adaptation techniques have contributed to the success of deep learning. Leveraging knowledge from an auxiliary source domain for learning in labeled data-scarce target domain is fundamental to domain adaptation. While these techniques result in increasing accuracy, the adaptation process, particularly the knowledge leveraged from the source domain, remains unclear. This paper proposes an explainable by design supervised domain adaptation framework - XSDA-Net. We integrate a case-based reasoning mechanism into the XSDA-Net to explain the prediction of a test instance in terms of similar-looking regions in the source and target train images. We empirically demonstrate the utility of the proposed framework by curating the domain adaptation settings on datasets popularly known to exhibit part-based explainability.

Conformal Prediction with Temporal Quantile Adjustments

We develop Temporal Quantile Adjustment (TQA), a general method to construct efficient and valid prediction intervals (PIs) for regression on cross-sectional time series data. Such data is common in many domains, including econometrics and healthcare. A canonical example in healthcare is predicting patient outcomes using physiological time-series data, where a population of patients composes a cross-section. Reliable PI estimators in this setting must address two distinct notions of coverage: cross-sectional coverage across a cross-sectional slice, and longitudinal coverage along the temporal dimension for each time series. Recent works have explored adapting Conformal Prediction (CP) to obtain PIs in the time series context. However, none handles both notions of coverage simultaneously. CP methods typically query a pre-specified quantile from the distribution of nonconformity scores on a calibration set. TQA adjusts the quantile to query in CP at each time $t$, accounting for both cross-sectional and longitudinal coverage in a theoretically-grounded manner. The post-hoc nature of TQA facilitates its use as a general wrapper around any time series regression model. We validate TQA's performance through extensive experimentation: TQA generally obtains efficient PIs and improves longitudinal coverage while preserving cross-sectional coverage.

Towards Explanation for Unsupervised Graph-Level Representation Learning

Due to the superior performance of Graph Neural Networks (GNNs) in various domains, there is an increasing interest in the GNN explanation problem "\emph{which fraction of the input graph is the most crucial to decide the model's decision?}" Existing explanation methods focus on the supervised settings, \eg, node classification and graph classification, while the explanation for unsupervised graph-level representation learning is still unexplored. The opaqueness of the graph representations may lead to unexpected risks when deployed for high-stake decision-making scenarios. In this paper, we advance the Information Bottleneck principle (IB) to tackle the proposed explanation problem for unsupervised graph representations, which leads to a novel principle, \textit{Unsupervised Subgraph Information Bottleneck} (USIB). We also theoretically analyze the connection between graph representations and explanatory subgraphs on the label space, which reveals that the expressiveness and robustness of representations benefit the fidelity of explanatory subgraphs. Experimental results on both synthetic and real-world datasets demonstrate the superiority of our developed explainer and the validity of our theoretical analysis.

FIND:Explainable Framework for Meta-learning

Meta-learning is used to efficiently enable the automatic selection of machine learning models by combining data and prior knowledge. Since the traditional meta-learning technique lacks explainability, as well as shortcomings in terms of transparency and fairness, achieving explainability for meta-learning is crucial. This paper proposes FIND, an interpretable meta-learning framework that not only can explain the recommendation results of meta-learning algorithm selection, but also provide a more complete and accurate explanation of the recommendation algorithm's performance on specific datasets combined with business scenarios. The validity and correctness of this framework have been demonstrated by extensive experiments.

BayesPCN: A Continually Learnable Predictive Coding Associative Memory

Associative memory plays an important role in human intelligence and its mechanisms have been linked to attention in machine learning. While the machine learning community's interest in associative memories has recently been rekindled, most work has focused on memory recall ($read$) over memory learning ($write$). In this paper, we present BayesPCN, a hierarchical associative memory capable of performing continual one-shot memory writes without meta-learning. Moreover, BayesPCN is able to gradually forget past observations ($forget$) to free its memory. Experiments show that BayesPCN can recall corrupted i.i.d. high-dimensional data observed hundreds of "timesteps" ago without a significant drop in recall ability compared to the state-of-the-art offline-learned associative memory models.

Cross Reconstruction Transformer for Self-Supervised Time Series Representation Learning

Unsupervised/self-supervised representation learning in time series is critical since labeled samples are usually scarce in real-world scenarios. Existing approaches mainly leverage the contrastive learning framework, which automatically learns to understand the similar and dissimilar data pairs. Nevertheless, they are restricted to the prior knowledge of constructing pairs, cumbersome sampling policy, and unstable performances when encountering sampling bias. Also, few works have focused on effectively modeling across temporal-spectral relations to extend the capacity of representations. In this paper, we aim at learning representations for time series from a new perspective and propose Cross Reconstruction Transformer (CRT) to solve the aforementioned problems in a unified way. CRT achieves time series representation learning through a cross-domain dropping-reconstruction task. Specifically, we transform time series into the frequency domain and randomly drop certain parts in both time and frequency domains. Dropping can maximally preserve the global context compared to cropping and masking. Then a transformer architecture is utilized to adequately capture the cross-domain correlations between temporal and spectral information through reconstructing data in both domains, which is called Dropped Temporal-Spectral Modeling. To discriminate the representations in global latent space, we propose Instance Discrimination Constraint to reduce the mutual information between different time series and sharpen the decision boundaries. Additionally, we propose a specified curriculum learning strategy to optimize the CRT, which progressively increases the dropping ratio in the training process.

CertiFair: A Framework for Certified Global Fairness of Neural Networks

We consider the problem of whether a Neural Network (NN) model satisfies global individual fairness. Individual Fairness suggests that similar individuals with respect to a certain task are to be treated similarly by the decision model. In this work, we have two main objectives. The first is to construct a verifier which checks whether the fairness property holds for a given NN in a classification task or provide a counterexample if it is violated, i.e., the model is fair if all similar individuals are classified the same, and unfair if a pair of similar individuals are classified differently. To that end, We construct a sound and complete verifier that verifies global individual fairness properties of ReLU NN classifiers using distance-based similarity metrics. The second objective of this paper is to provide a method for training provably fair NN classifiers from unfair (biased) data. We propose a fairness loss that can be used during training to enforce fair outcomes for similar individuals. We then provide provable bounds on the fairness of the resulting NN. We run experiments on commonly used fairness datasets that are publicly available and we show that global individual fairness can be improved by 96 % without significant drop in test accuracy.

On Jointly Optimizing Partial Offloading and SFC Mapping: A Cooperative Dual-agent Deep Reinforcement Learning Approach

Multi-access edge computing (MEC) and network function virtualization (NFV) are promising technologies to support emerging IoT applications, especially those computation-intensive. In NFV-enabled MEC environment, service function chain (SFC), i.e., a set of ordered virtual network functions (VNFs), can be mapped on MEC servers. Mobile devices (MDs) can offload computation-intensive applications, which can be represented by SFCs, fully or partially to MEC servers for remote execution. This paper studies the partial offloading and SFC mapping joint optimization (POSMJO) problem in an NFV-enabled MEC system, where an incoming task can be partitioned into two parts, one for local execution and the other for remote execution. The objective is to minimize the average cost in the long term which is a combination of execution delay, MD's energy consumption, and usage charge for edge computing. This problem consists of two closely related decision-making steps, namely task partition and VNF placement, which is highly complex and quite challenging. To address this, we propose a cooperative dual-agent deep reinforcement learning (CDADRL) algorithm, where we design a framework enabling interaction between two agents. Simulation results show that the proposed algorithm outperforms three combinations of deep reinforcement learning algorithms in terms of cumulative and average episodic rewards and it overweighs a number of baseline algorithms with respect to execution delay, energy consumption, and usage charge.

Anomaly Detection for Multivariate Time Series on Large-scale Fluid Handling Plant Using Two-stage Autoencoder

This paper focuses on anomaly detection for multivariate time series data in large-scale fluid handling plants with dynamic components, such as power generation, water treatment, and chemical plants, where signals from various physical phenomena are observed simultaneously. In these plants, the need for anomaly detection techniques is increasing in order to reduce the cost of operation and maintenance, in view of a decline in the number of skilled engineers and a shortage of manpower. However, considering the complex behavior of high-dimensional signals and the demand for interpretability, the techniques constitute a major challenge. We introduce a Two-Stage AutoEncoder (TSAE) as an anomaly detection method suitable for such plants. This is a simple autoencoder architecture that makes anomaly detection more interpretable and more accurate, in which based on the premise that plant signals can be separated into two behaviors that have almost no correlation with each other, the signals are separated into long-term and short-term components in a stepwise manner, and the two components are trained independently to improve the inference capability for normal signals. Through experiments on two publicly available datasets of water treatment systems, we have confirmed the high detection performance, the validity of the premise, and that the model behavior was as intended, i.e., the technical effectiveness of TSAE.

KERPLE: Kernelized Relative Positional Embedding for Length Extrapolation

Relative positional embeddings (RPE) have received considerable attention since RPEs effectively model the relative distance among tokens and enable length extrapolation. We propose KERPLE, a framework that generalizes relative position embedding for extrapolation by kernelizing positional differences. We achieve this goal using conditionally positive definite (CPD) kernels, a class of functions known for generalizing distance metrics. To maintain the inner product interpretation of self-attention, we show that a CPD kernel can be transformed into a PD kernel by adding a constant offset. This offset is implicitly absorbed in the Softmax normalization during self-attention. The diversity of CPD kernels allows us to derive various RPEs that enable length extrapolation in a principled way. Experiments demonstrate that the logarithmic variant achieves excellent extrapolation performance on three large language modeling datasets.

Robust Expected Information Gain for Optimal Bayesian Experimental Design Using Ambiguity Sets

The ranking of experiments by expected information gain (EIG) in Bayesian experimental design is sensitive to changes in the model's prior distribution, and the approximation of EIG yielded by sampling will have errors similar to the use of a perturbed prior. We define and analyze \emph{robust expected information gain} (REIG), a modification of the objective in EIG maximization by minimizing an affine relaxation of EIG over an ambiguity set of distributions that are close to the original prior in KL-divergence. We show that, when combined with a sampling-based approach to estimating EIG, REIG corresponds to a `log-sum-exp' stabilization of the samples used to estimate EIG, meaning that it can be efficiently implemented in practice. Numerical tests combining REIG with variational nested Monte Carlo (VNMC), adaptive contrastive estimation (ACE) and mutual information neural estimation (MINE) suggest that in practice REIG also compensates for the variability of under-sampled estimators.

Can Foundation Models Wrangle Your Data?

Foundation Models (FMs) are models trained on large corpora of data that, at very large scale, can generalize to new tasks without any task-specific finetuning. As these models continue to grow in size, innovations continue to push the boundaries of what these models can do on language and image tasks. This paper aims to understand an underexplored area of FMs: classical data tasks like cleaning and integration. As a proof-of-concept, we cast three data cleaning and integration tasks as prompting tasks and evaluate the performance of FMs on these tasks. We find that large FMs generalize and achieve SoTA performance on data cleaning and integration tasks, even though they are not trained for these data tasks. We identify specific research challenges and opportunities that these models present, including challenges with private and temporal data, and opportunities to make data driven systems more accessible to non-experts. We make our code and experiments publicly available at: https://github.com/HazyResearch/fm_data_tasks.

Sparse Infinite Random Feature Latent Variable Modeling

We propose a non-linear, Bayesian non-parametric latent variable model where the latent space is assumed to be sparse and infinite dimensional a priori using an Indian buffet process prior. A posteriori, the number of instantiated dimensions in the latent space is guaranteed to be finite. The purpose of placing the Indian buffet process on the latent variables is to: 1.) Automatically and probabilistically select the number of latent dimensions. 2.) Impose sparsity in the latent space, where the Indian buffet process will select which elements are exactly zero. Our proposed model allows for sparse, non-linear latent variable modeling where the number of latent dimensions is selected automatically. Inference is made tractable using the random Fourier approximation and we can easily implement posterior inference through Markov chain Monte Carlo sampling. This approach is amenable to many observation models beyond the Gaussian setting. We demonstrate the utility of our method on a variety of synthetic, biological and text datasets and show that we can obtain superior test set performance compared to previous latent variable models.

Data Augmentation for Compositional Data: Advancing Predictive Models of the Microbiome

Data augmentation plays a key role in modern machine learning pipelines. While numerous augmentation strategies have been studied in the context of computer vision and natural language processing, less is known for other data modalities. Our work extends the success of data augmentation to compositional data, i.e., simplex-valued data, which is of particular interest in the context of the human microbiome. Drawing on key principles from compositional data analysis, such as the Aitchison geometry of the simplex and subcompositions, we define novel augmentation strategies for this data modality. Incorporating our data augmentations into standard supervised learning pipelines results in consistent performance gains across a wide range of standard benchmark datasets. In particular, we set a new state-of-the-art for key disease prediction tasks including colorectal cancer, type 2 diabetes, and Crohn's disease. In addition, our data augmentations enable us to define a novel contrastive learning model, which improves on previous representation learning approaches for microbiome compositional data. Our code is available at https://github.com/cunningham-lab/AugCoDa.

Manifold-aligned Neighbor Embedding

In this paper, we introduce a neighbor embedding framework for manifold alignment. We demonstrate the efficacy of the framework using a manifold-aligned version of the uniform manifold approximation and projection algorithm. We show that our algorithm can learn an aligned manifold that is visually competitive to embedding of the whole dataset.

Minimal Explanations for Neural Network Predictions

Explaining neural network predictions is known to be a challenging problem. In this paper, we propose a novel approach which can be effectively exploited, either in isolation or in combination with other methods, to enhance the interpretability of neural model predictions. For a given input to a trained neural model, our aim is to compute a smallest set of input features so that the model prediction changes when these features are disregarded by setting them to an uninformative baseline value. While computing such minimal explanations is computationally intractable in general for fully-connected neural networks, we show that the problem becomes solvable in polynomial time by a greedy algorithm under mild assumptions on the network's activation functions. We then show that our tractability result extends seamlessly to more advanced neural architectures such as convolutional and graph neural networks. We conduct experiments to showcase the capability of our method for identifying the input features that are essential to the model's prediction.

Estimating the frame potential of large-scale quantum circuit sampling using tensor networks up to 50 qubits

We develop numerical protocols for estimating the frame potential, the 2-norm distance between a given ensemble and the exact Haar randomness, using the \texttt{QTensor} platform. Our tensor-network-based algorithm has polynomial complexity for shallow circuits and is high performing using CPU and GPU parallelism. We apply the above methods to two problems: the Brown-Susskind conjecture, with local and parallel random circuits in terms of the Haar distance and the approximate $k$-design properties of the hardware efficient ans{\"a}tze in quantum machine learning, which induce the barren plateau problem. We estimate frame potentials with these ensembles up to 50 qubits and $k=5$, examine the Haar distance of the hardware-efficient ans{\"a}tze, and verify the Brown-Susskind conjecture numerically. Our work shows that large-scale tensor network simulations could provide important hints toward open problems in quantum information science.

Breaking the $\sqrt{T}$ Barrier: Instance-Independent Logarithmic Regret in Stochastic Contextual Linear Bandits

We prove an instance independent (poly) logarithmic regret for stochastic contextual bandits with linear payoff. Previously, in \cite{chu2011contextual}, a lower bound of $\mathcal{O}(\sqrt{T})$ is shown for the contextual linear bandit problem with arbitrary (adversarily chosen) contexts. In this paper, we show that stochastic contexts indeed help to reduce the regret from $\sqrt{T}$ to $\polylog(T)$. We propose Low Regret Stochastic Contextual Bandits (\texttt{LR-SCB}), which takes advantage of the stochastic contexts and performs parameter estimation (in $\ell_2$ norm) and regret minimization simultaneously. \texttt{LR-SCB} works in epochs, where the parameter estimation of the previous epoch is used to reduce the regret of the current epoch. The (poly) logarithmic regret of \texttt{LR-SCB} stems from two crucial facts: (a) the application of a norm adaptive algorithm to exploit the parameter estimation and (b) an analysis of the shifted linear contextual bandit algorithm, showing that shifting results in increasing regret. We have also shown experimentally that stochastic contexts indeed incurs a regret that scales with $\polylog(T)$.

Let the Model Decide its Curriculum for Multitask Learning

Curriculum learning strategies in prior multi-task learning approaches arrange datasets in a difficulty hierarchy either based on human perception or by exhaustively searching the optimal arrangement. However, human perception of difficulty may not always correlate well with machine interpretation leading to poor performance and exhaustive search is computationally expensive. Addressing these concerns, we propose two classes of techniques to arrange training instances into a learning curriculum based on difficulty scores computed via model-based approaches. The two classes i.e Dataset-level and Instance-level differ in granularity of arrangement. Through comprehensive experiments with 12 datasets, we show that instance-level and dataset-level techniques result in strong representations as they lead to an average performance improvement of 4.17% and 3.15% over their respective baselines. Furthermore, we find that most of this improvement comes from correctly answering the difficult instances, implying a greater efficacy of our techniques on difficult tasks.