title
stringlengths 7
246
| abstract
stringlengths 3
3.31k
|
|---|---|
Optimal Algorithms for Stochastic Multi-Level Compositional Optimization | In this paper, we investigate the problem of stochastic multi-level compositional optimization, where the objective function is a composition of multiple smooth but possibly non-convex functions. Existing methods for solving this problem either suffer from sub-optimal sample complexities or need a huge batch size. To address this limitation, we propose a Stochastic Multi-level Variance Reduction method (SMVR), which achieves the optimal sample complexity of $\mathcal{O}\left(1 / \epsilon^{3}\right)$ to find an $\epsilon$-stationary point for non-convex objectives. Furthermore, when the objective function satisfies the convexity or Polyak-{\L}ojasiewicz (PL) condition, we propose a stage-wise variant of SMVR and improve the sample complexity to $\mathcal{O}\left(1 / \epsilon^{2}\right)$ for convex functions or $\mathcal{O}\left(1 /(\mu\epsilon)\right)$ for non-convex functions satisfying the $\mu$-PL condition. The latter result implies the same complexity for $\mu$-strongly convex functions. To make use of adaptive learning rates, we also develop Adaptive SMVR, which achieves the same optimal complexities but converges faster in practice. All our complexities match the lower bounds not only in terms of $\epsilon$ but also in terms of $\mu$ (for PL or strongly convex functions), without using a large batch size in each iteration. |
Enhancing Identification of Causal Effects by Pruning | Causal models communicate our assumptions about causes and effects in real-world phe- nomena. Often the interest lies in the identification of the effect of an action which means deriving an expression from the observed probability distribution for the interventional distribution resulting from the action. In many cases an identifiability algorithm may return a complicated expression that contains variables that are in fact unnecessary. In practice this can lead to additional computational burden and increased bias or inefficiency of estimates when dealing with measurement error or missing data. We present graphical criteria to detect variables which are redundant in identifying causal effects. We also provide an improved version of a well-known identifiability algorithm that implements these criteria. |
Spectral Clustering Revisited: Information Hidden in the Fiedler Vector | We are interested in the clustering problem on graphs: it is known that if there are two underlying clusters, then the signs of the eigenvector corresponding to the second largest eigenvalue of the adjacency matrix can reliably reconstruct the two clusters. We argue that the vertices for which the eigenvector has the largest and the smallest entries, respectively, are unusually strongly connected to their own cluster and more reliably classified than the rest. This can be regarded as a discrete version of the Hot Spots conjecture and should be useful in applications. We give a rigorous proof for the stochastic block model and several examples. |
Neural Policy Gradient Methods: Global Optimality and Rates of Convergence | Policy gradient methods with actor-critic schemes demonstrate tremendous empirical successes, especially when the actors and critics are parameterized by neural networks. However, it remains less clear whether such "neural" policy gradient methods converge to globally optimal policies and whether they even converge at all. We answer both the questions affirmatively in the overparameterized regime. In detail, we prove that neural natural policy gradient converges to a globally optimal policy at a sublinear rate. Also, we show that neural vanilla policy gradient converges sublinearly to a stationary point. Meanwhile, by relating the suboptimality of the stationary points to the representation power of neural actor and critic classes, we prove the global optimality of all stationary points under mild regularity conditions. Particularly, we show that a key to the global optimality and convergence is the "compatibility" between the actor and critic, which is ensured by sharing neural architectures and random initializations across the actor and critic. To the best of our knowledge, our analysis establishes the first global optimality and convergence guarantees for neural policy gradient methods. |
Decorrelated Clustering with Data Selection Bias | Most of existing clustering algorithms are proposed without considering the selection bias in data. In many real applications, however, one cannot guarantee the data is unbiased. Selection bias might bring the unexpected correlation between features and ignoring those unexpected correlations will hurt the performance of clustering algorithms. Therefore, how to remove those unexpected correlations induced by selection bias is extremely important yet largely unexplored for clustering. In this paper, we propose a novel Decorrelation regularized K-Means algorithm (DCKM) for clustering with data selection bias. Specifically, the decorrelation regularizer aims to learn the global sample weights which are capable of balancing the sample distribution, so as to remove unexpected correlations among features. Meanwhile, the learned weights are combined with k-means, which makes the reweighted k-means cluster on the inherent data distribution without unexpected correlation influence. Moreover, we derive the updating rules to effectively infer the parameters in DCKM. Extensive experiments results on real world datasets well demonstrate that our DCKM algorithm achieves significant performance gains, indicating the necessity of removing unexpected feature correlations induced by selection bias when clustering. |
Multi Sense Embeddings from Topic Models | Distributed word embeddings have yielded state-of-the-art performance in many NLP tasks, mainly due to their success in capturing useful semantic information. These representations assign only a single vector to each word whereas a large number of words are polysemous (i.e., have multiple meanings). In this work, we approach this critical problem in lexical semantics, namely that of representing various senses of polysemous words in vector spaces. We propose a topic modeling based skip-gram approach for learning multi-prototype word embeddings. We also introduce a method to prune the embeddings determined by the probabilistic representation of the word in each topic. We use our embeddings to show that they can capture the context and word similarity strongly and outperform various state-of-the-art implementations. |
Network of Evolvable Neural Units: Evolving to Learn at a Synaptic Level | Although Deep Neural Networks have seen great success in recent years through various changes in overall architectures and optimization strategies, their fundamental underlying design remains largely unchanged. Computational neuroscience on the other hand provides more biologically realistic models of neural processing mechanisms, but they are still high level abstractions of the actual experimentally observed behaviour. Here a model is proposed that bridges Neuroscience, Machine Learning and Evolutionary Algorithms to evolve individual soma and synaptic compartment models of neurons in a scalable manner. Instead of attempting to manually derive models for all the observed complexity and diversity in neural processing, we propose an Evolvable Neural Unit (ENU) that can approximate the function of each individual neuron and synapse. We demonstrate that this type of unit can be evolved to mimic Integrate-And-Fire neurons and synaptic Spike-Timing-Dependent Plasticity. Additionally, by constructing a new type of neural network where each synapse and neuron is such an evolvable neural unit, we show it is possible to evolve an agent capable of learning to solve a T-maze environment task. This network independently discovers spiking dynamics and reinforcement type learning rules, opening up a new path towards biologically inspired artificial intelligence. |
Recurrent Autoregressive Networks for Online Multi-Object Tracking | The main challenge of online multi-object tracking is to reliably associate object trajectories with detections in each video frame based on their tracking history. In this work, we propose the Recurrent Autoregressive Network (RAN), a temporal generative modeling framework to characterize the appearance and motion dynamics of multiple objects over time. The RAN couples an external memory and an internal memory. The external memory explicitly stores previous inputs of each trajectory in a time window, while the internal memory learns to summarize long-term tracking history and associate detections by processing the external memory. We conduct experiments on the MOT 2015 and 2016 datasets to demonstrate the robustness of our tracking method in highly crowded and occluded scenes. Our method achieves top-ranked results on the two benchmarks. |
HarperValleyBank: A Domain-Specific Spoken Dialog Corpus | We introduce HarperValleyBank, a free, public domain spoken dialog corpus. The data simulate simple consumer banking interactions, containing about 23 hours of audio from 1,446 human-human conversations between 59 unique speakers. We selected intents and utterance templates to allow realistic variation while controlling overall task complexity and limiting vocabulary size to about 700 unique words. We provide audio data along with transcripts and annotations for speaker identity, caller intent, dialog actions, and emotional valence. The data size and domain specificity makes for quick transcription experiments with modern end-to-end neural approaches. Further, we provide baselines for representation learning, adapting recent work to embed waveforms for downstream prediction tasks. Our experiments show that tasks using our annotations are sensitive to both the model choice and corpus size. |
Learning to Rank under Multinomial Logit Choice | Learning the optimal ordering of content is an important challenge in website design. The learning to rank (LTR) framework models this problem as a sequential problem of selecting lists of content and observing where users decide to click. Most previous work on LTR assumes that the user considers each item in the list in isolation, and makes binary choices to click or not on each. We introduce a multinomial logit (MNL) choice model to the LTR framework, which captures the behaviour of users who consider the ordered list of items as a whole and make a single choice among all the items and a no-click option. Under the MNL model, the user favours items which are either inherently more attractive, or placed in a preferable position within the list. We propose upper confidence bound algorithms to minimise regret in two settings - where the position dependent parameters are known, and unknown. We present theoretical analysis leading to an $\Omega(\sqrt{T})$ lower bound for the problem, an $\tilde{O}(\sqrt{T})$ upper bound on regret for the known parameter version. Our analyses are based on tight new concentration results for Geometric random variables, and novel functional inequalities for maximum likelihood estimators computed on discrete data. |
Learning Improved Representations by Transferring Incomplete Evidence Across Heterogeneous Tasks | Acquiring ground truth labels for unlabelled data can be a costly procedure, since it often requires manual labour that is error-prone. Consequently, the available amount of labelled data is increasingly reduced due to the limitations of manual data labelling. It is possible to increase the amount of labelled data samples by performing automated labelling or crowd-sourcing the annotation procedure. However, they often introduce noise or uncertainty in the labelset, that leads to decreased performance of supervised deep learning methods. On the other hand, weak supervision methods remain robust during noisy labelsets or can be effective even with low amounts of labelled data. In this paper we evaluate the effectiveness of a representation learning method that uses external categorical evidence called "Evidence Transfer", against low amount of corresponding evidence termed as incomplete evidence. Evidence transfer is a robust solution against external unknown categorical evidence that can introduce noise or uncertainty. In our experimental evaluation, evidence transfer proves to be effective and robust against different levels of incompleteness, for two types of incomplete evidence. |
Extended T: Learning with Mixed Closed-set and Open-set Noisy Labels | The label noise transition matrix $T$, reflecting the probabilities that true labels flip into noisy ones, is of vital importance to model label noise and design statistically consistent classifiers. The traditional transition matrix is limited to model closed-set label noise, where noisy training data has true class labels within the noisy label set. It is unfitted to employ such a transition matrix to model open-set label noise, where some true class labels are outside the noisy label set. Thus when considering a more realistic situation, i.e., both closed-set and open-set label noise occurs, existing methods will undesirably give biased solutions. Besides, the traditional transition matrix is limited to model instance-independent label noise, which may not perform well in practice. In this paper, we focus on learning under the mixed closed-set and open-set label noise. We address the aforementioned issues by extending the traditional transition matrix to be able to model mixed label noise, and further to the cluster-dependent transition matrix to better approximate the instance-dependent label noise in real-world applications. We term the proposed transition matrix as the cluster-dependent extended transition matrix. An unbiased estimator (i.e., extended $T$-estimator) has been designed to estimate the cluster-dependent extended transition matrix by only exploiting the noisy data. Comprehensive synthetic and real experiments validate that our method can better model the mixed label noise, following its more robust performance than the prior state-of-the-art label-noise learning methods. |
Dynamic RAN Slicing for Service-Oriented Vehicular Networks via Constrained Learning | In this paper, we investigate a radio access network (RAN) slicing problem for Internet of vehicles (IoV) services with different quality of service (QoS) requirements, in which multiple logically-isolated slices are constructed on a common roadside network infrastructure. A dynamic RAN slicing framework is presented to dynamically allocate radio spectrum and computing resource, and distribute computation workloads for the slices. To obtain an optimal RAN slicing policy for accommodating the spatial-temporal dynamics of vehicle traffic density, we first formulate a constrained RAN slicing problem with the objective to minimize long-term system cost. This problem cannot be directly solved by traditional reinforcement learning (RL) algorithms due to complicated coupled constraints among decisions. Therefore, we decouple the problem into a resource allocation subproblem and a workload distribution subproblem, and propose a two-layer constrained RL algorithm, named Resource Allocation and Workload diStribution (RAWS) to solve them. Specifically, an outer layer first makes the resource allocation decision via an RL algorithm, and then an inner layer makes the workload distribution decision via an optimization subroutine. Extensive trace-driven simulations show that the RAWS effectively reduces the system cost while satisfying QoS requirements with a high probability, as compared with benchmarks. |
Exploring Vision Transformers for Fine-grained Classification | Existing computer vision research in categorization struggles with fine-grained attributes recognition due to the inherently high intra-class variances and low inter-class variances. SOTA methods tackle this challenge by locating the most informative image regions and rely on them to classify the complete image. The most recent work, Vision Transformer (ViT), shows its strong performance in both traditional and fine-grained classification tasks. In this work, we propose a multi-stage ViT framework for fine-grained image classification tasks, which localizes the informative image regions without requiring architectural changes using the inherent multi-head self-attention mechanism. We also introduce attention-guided augmentations for improving the model's capabilities. We demonstrate the value of our approach by experimenting with four popular fine-grained benchmarks: CUB-200-2011, Stanford Cars, Stanford Dogs, and FGVC7 Plant Pathology. We also prove our model's interpretability via qualitative results. |
Falsification of Multiple Requirements for Cyber-Physical Systems Using Online Generative Adversarial Networks and Multi-Armed Bandits | We consider the problem of falsifying safety requirements of Cyber-Physical Systems expressed in signal temporal logic (STL). This problem can be turned into an optimization problem via STL robustness functions. In this paper, our focus is in falsifying systems with multiple requirements. We propose to solve such conjunctive requirements using online generative adversarial networks (GANs) as test generators. Our main contribution is an algorithm which falsifies a conjunctive requirement $\varphi_1 \land \cdots \land \varphi_n$ by using a GAN for each requirement $\varphi_i$ separately. Using ideas from multi-armed bandit algorithms, our algorithm only trains a single GAN at every step, which saves resources. Our experiments indicate that, in addition to saving resources, this multi-armed bandit algorithm can falsify requirements with fewer number of executions on the system under test when compared to (i) an algorithm training a single GAN for the complete conjunctive requirement and (ii) an algorithm always training $n$ GANs at each step. |
A review of mean-shift algorithms for clustering | A natural way to characterize the cluster structure of a dataset is by finding regions containing a high density of data. This can be done in a nonparametric way with a kernel density estimate, whose modes and hence clusters can be found using mean-shift algorithms. We describe the theory and practice behind clustering based on kernel density estimates and mean-shift algorithms. We discuss the blurring and non-blurring versions of mean-shift; theoretical results about mean-shift algorithms and Gaussian mixtures; relations with scale-space theory, spectral clustering and other algorithms; extensions to tracking, to manifold and graph data, and to manifold denoising; K-modes and Laplacian K-modes algorithms; acceleration strategies for large datasets; and applications to image segmentation, manifold denoising and multivalued regression. |
Random gradient extrapolation for distributed and stochastic optimization | In this paper, we consider a class of finite-sum convex optimization problems defined over a distributed multiagent network with $m$ agents connected to a central server. In particular, the objective function consists of the average of $m$ ($\ge 1$) smooth components associated with each network agent together with a strongly convex term. Our major contribution is to develop a new randomized incremental gradient algorithm, namely random gradient extrapolation method (RGEM), which does not require any exact gradient evaluation even for the initial point, but can achieve the optimal ${\cal O}(\log(1/\epsilon))$ complexity bound in terms of the total number of gradient evaluations of component functions to solve the finite-sum problems. Furthermore, we demonstrate that for stochastic finite-sum optimization problems, RGEM maintains the optimal ${\cal O}(1/\epsilon)$ complexity (up to a certain logarithmic factor) in terms of the number of stochastic gradient computations, but attains an ${\cal O}(\log(1/\epsilon))$ complexity in terms of communication rounds (each round involves only one agent). It is worth noting that the former bound is independent of the number of agents $m$, while the latter one only linearly depends on $m$ or even $\sqrt m$ for ill-conditioned problems. To the best of our knowledge, this is the first time that these complexity bounds have been obtained for distributed and stochastic optimization problems. Moreover, our algorithms were developed based on a novel dual perspective of Nesterov's accelerated gradient method. |
Machine learning approach for early detection of autism by combining questionnaire and home video screening | Existing screening tools for early detection of autism are expensive, cumbersome, time-intensive, and sometimes fall short in predictive value. In this work, we apply Machine Learning (ML) to gold standard clinical data obtained across thousands of children at risk for autism spectrum disorders to create a low-cost, quick, and easy to apply autism screening tool that performs as well or better than most widely used standardized instruments. This new tool combines two screening methods into a single assessment, one based on short, structured parent-report questionnaires and the other on tagging key behaviors from short, semi-structured home videos of children. To overcome the scarcity, sparsity, and imbalance of training data, we apply creative feature selection, feature engineering, and novel feature encoding techniques. We allow for inconclusive determination where appropriate in order to boost screening accuracy when conclusive. We demonstrate a significant accuracy improvement over standard screening tools in a clinical study sample of 162 children. |
Force2Vec: Parallel force-directed graph embedding | A graph embedding algorithm embeds a graph into a low-dimensional space such that the embedding preserves the inherent properties of the graph. While graph embedding is fundamentally related to graph visualization, prior work did not exploit this connection explicitly. We develop Force2Vec that uses force-directed graph layout models in a graph embedding setting with an aim to excel in both machine learning (ML) and visualization tasks. We make Force2Vec highly parallel by mapping its core computations to linear algebra and utilizing multiple levels of parallelism available in modern processors. The resultant algorithm is an order of magnitude faster than existing methods (43x faster than DeepWalk, on average) and can generate embeddings from graphs with billions of edges in a few hours. In comparison to existing methods, Force2Vec is better in graph visualization and performs comparably or better in ML tasks such as link prediction, node classification, and clustering. Source code is available at https://github.com/HipGraph/Force2Vec. |
One model Packs Thousands of Items with Recurrent Conditional Query Learning | Recent studies have revealed that neural combinatorial optimization (NCO) has advantages over conventional algorithms in many combinatorial optimization problems such as routing, but it is less efficient for more complicated optimization tasks such as packing which involves mutually conditioned action spaces. In this paper, we propose a Recurrent Conditional Query Learning (RCQL) method to solve both 2D and 3D packing problems. We first embed states by a recurrent encoder, and then adopt attention with conditional queries from previous actions. The conditional query mechanism fills the information gap between learning steps, which shapes the problem as a Markov decision process. Benefiting from the recurrence, a single RCQL model is capable of handling different sizes of packing problems. Experiment results show that RCQL can effectively learn strong heuristics for offline and online strip packing problems (SPPs), outperforming a wide range of baselines in space utilization ratio. RCQL reduces the average bin gap ratio by 1.83% in offline 2D 40-box cases and 7.84% in 3D cases compared with state-of-the-art methods. Meanwhile, our method also achieves 5.64% higher space utilization ratio for SPPs with 1000 items than the state of the art. |
Detecting Dynamic Community Structure in Functional Brain Networks Across Individuals: A Multilayer Approach | We present a unified statistical framework for characterizing community structure of brain functional networks that captures variation across individuals and evolution over time. Existing methods for community detection focus only on single-subject analysis of dynamic networks; while recent extensions to multiple-subjects analysis are limited to static networks. To overcome these limitations, we propose a multi-subject, Markov-switching stochastic block model (MSS-SBM) to identify state-related changes in brain community organization over a group of individuals. We first formulate a multilayer extension of SBM to describe the time-dependent, multi-subject brain networks. We develop a novel procedure for fitting the multilayer SBM that builds on multislice modularity maximization which can uncover a common community partition of all layers (subjects) simultaneously. By augmenting with a dynamic Markov switching process, our proposed method is able to capture a set of distinct, recurring temporal states with respect to inter-community interactions over subjects and the change points between them. Simulation shows accurate community recovery and tracking of dynamic community regimes over multilayer networks by the MSS-SBM. Application to task fMRI reveals meaningful non-assortative brain community motifs, e.g., core-periphery structure at the group level, that are associated with language comprehension and motor functions suggesting their putative role in complex information integration. Our approach detected dynamic reconfiguration of modular connectivity elicited by varying task demands and identified unique profiles of intra and inter-community connectivity across different task conditions. The proposed multilayer network representation provides a principled way of detecting synchronous, dynamic modularity in brain networks across subjects. |
Online Baum-Welch algorithm for Hierarchical Imitation Learning | The options framework for hierarchical reinforcement learning has increased its popularity in recent years and has made improvements in tackling the scalability problem in reinforcement learning. Yet, most of these recent successes are linked with a proper options initialization or discovery. When an expert is available, the options discovery problem can be addressed by learning an options-type hierarchical policy directly from expert demonstrations. This problem is referred to as hierarchical imitation learning and can be handled as an inference problem in a Hidden Markov Model, which is done via an Expectation-Maximization type algorithm. In this work, we propose a novel online algorithm to perform hierarchical imitation learning in the options framework. Further, we discuss the benefits of such an algorithm and compare it with its batch version in classical reinforcement learning benchmarks. We show that this approach works well in both discrete and continuous environments and, under certain conditions, it outperforms the batch version. |
Not All Images are Worth 16x16 Words: Dynamic Transformers for Efficient Image Recognition | Vision Transformers (ViT) have achieved remarkable success in large-scale image recognition. They split every 2D image into a fixed number of patches, each of which is treated as a token. Generally, representing an image with more tokens would lead to higher prediction accuracy, while it also results in drastically increased computational cost. To achieve a decent trade-off between accuracy and speed, the number of tokens is empirically set to 16x16 or 14x14. In this paper, we argue that every image has its own characteristics, and ideally the token number should be conditioned on each individual input. In fact, we have observed that there exist a considerable number of "easy" images which can be accurately predicted with a mere number of 4x4 tokens, while only a small fraction of "hard" ones need a finer representation. Inspired by this phenomenon, we propose a Dynamic Transformer to automatically configure a proper number of tokens for each input image. This is achieved by cascading multiple Transformers with increasing numbers of tokens, which are sequentially activated in an adaptive fashion at test time, i.e., the inference is terminated once a sufficiently confident prediction is produced. We further design efficient feature reuse and relationship reuse mechanisms across different components of the Dynamic Transformer to reduce redundant computations. Extensive empirical results on ImageNet, CIFAR-10, and CIFAR-100 demonstrate that our method significantly outperforms the competitive baselines in terms of both theoretical computational efficiency and practical inference speed. Code and pre-trained models (based on PyTorch and MindSpore) are available at https://github.com/blackfeather-wang/Dynamic-Vision-Transformer and https://github.com/blackfeather-wang/Dynamic-Vision-Transformer-MindSpore. |
Information Directed Sampling for Stochastic Bandits with Graph Feedback | We consider stochastic multi-armed bandit problems with graph feedback, where the decision maker is allowed to observe the neighboring actions of the chosen action. We allow the graph structure to vary with time and consider both deterministic and Erd\H{o}s-R\'enyi random graph models. For such a graph feedback model, we first present a novel analysis of Thompson sampling that leads to tighter performance bound than existing work. Next, we propose new Information Directed Sampling based policies that are graph-aware in their decision making. Under the deterministic graph case, we establish a Bayesian regret bound for the proposed policies that scales with the clique cover number of the graph instead of the number of actions. Under the random graph case, we provide a Bayesian regret bound for the proposed policies that scales with the ratio of the number of actions over the expected number of observations per iteration. To the best of our knowledge, this is the first analytical result for stochastic bandits with random graph feedback. Finally, using numerical evaluations, we demonstrate that our proposed IDS policies outperform existing approaches, including adaptions of upper confidence bound, $\epsilon$-greedy and Exp3 algorithms. |
Learning Deep Latent Subspaces for Image Denoising | Heterogeneity exists in most camera images. This heterogeneity manifests itself across the image space as varied Moire ringing, motion-blur, color-bleaching or lens based projection distortions. Moreover, combinations of these image artifacts can be present in small or large pixel neighborhoods, within an acquired image. Current camera image processing pipelines, including deep trained versions, tend to rectify the issue applying a single filter that is homogeneously applied to the entire image. This is also particularly true when an encoder-decoder type deep architecture is trained for the task. In this paper, we present a structured deep learning model that solves the heterogeneous image artifact filtering problem. We call our deep trained model the Patch Subspace Variational Autoencoder (PS-VAE) for Camera ISP. PS-VAE does not necessarily assume uniform image distortion levels nor similar artifact types within the image. Rather, our model attempts to learn to cluster different patches extracted from images into artifact type and distortion levels, within multiple latent subspaces (e.g. Moire ringing artifacts are often a higher dimensional latent distortion than a Gaussian motion blur artifact). Each image's patches are encoded into soft-clusters in their appropriate latent sub-space, using a prior mixture model. The decoders of the PS-VAE are also trained in an unsupervised manner for each of the image patches in each soft-cluster. Our experimental results demonstrates the flexibility and performance that one can achieve through improved heterogeneous filtering. We compare our results to a conventional one-encoder-one-decoder architecture. |
A gray-box approach for curriculum learning | Curriculum learning is often employed in deep reinforcement learning to let the agent progress more quickly towards better behaviors. Numerical methods for curriculum learning in the literature provides only initial heuristic solutions, with little to no guarantee on their quality. We define a new gray-box function that, including a suitable scheduling problem, can be effectively used to reformulate the curriculum learning problem. We propose different efficient numerical methods to address this gray-box reformulation. Preliminary numerical results on a benchmark task in the curriculum learning literature show the viability of the proposed approach. |
Sharper Rates and Flexible Framework for Nonconvex SGD with Client and Data Sampling | We revisit the classical problem of finding an approximately stationary point of the average of $n$ smooth and possibly nonconvex functions. The optimal complexity of stochastic first-order methods in terms of the number of gradient evaluations of individual functions is $\mathcal{O}\left(n + n^{1/2}\varepsilon^{-1}\right)$, attained by the optimal SGD methods $\small\sf\color{green}{SPIDER}$(arXiv:1807.01695) and $\small\sf\color{green}{PAGE}$(arXiv:2008.10898), for example, where $\varepsilon$ is the error tolerance. However, i) the big-$\mathcal{O}$ notation hides crucial dependencies on the smoothness constants associated with the functions, and ii) the rates and theory in these methods assume simplistic sampling mechanisms that do not offer any flexibility. In this work we remedy the situation. First, we generalize the $\small\sf\color{green}{PAGE}$ algorithm so that it can provably work with virtually any (unbiased) sampling mechanism. This is particularly useful in federated learning, as it allows us to construct and better understand the impact of various combinations of client and data sampling strategies. Second, our analysis is sharper as we make explicit use of certain novel inequalities that capture the intricate interplay between the smoothness constants and the sampling procedure. Indeed, our analysis is better even for the simple sampling procedure analyzed in the $\small\sf\color{green}{PAGE}$ paper. However, this already improved bound can be further sharpened by a different sampling scheme which we propose. In summary, we provide the most general and most accurate analysis of optimal SGD in the smooth nonconvex regime. Finally, our theoretical findings are supposed with carefully designed experiments. |
Bioinspired random projections for robust, sparse classification | Inspired by the use of random projections in biological sensing systems, we present a new algorithm for processing data in classification problems. This is based on observations of the human brain and the fruit fly's olfactory system and involves randomly projecting data into a space of greatly increased dimension before applying a cap operation to truncate the smaller entries. This leads to an algorithm that achieves a sparse representation with minimal loss in classification accuracy and is also more robust in the sense that classification accuracy is improved when noise is added to the data. This is demonstrated with numerical experiments, which supplement theoretical results demonstrating that the resulting signal transform is continuous and invertible, in an appropriate sense. |
Mathematical Models of Overparameterized Neural Networks | Deep learning has received considerable empirical successes in recent years. However, while many ad hoc tricks have been discovered by practitioners, until recently, there has been a lack of theoretical understanding for tricks invented in the deep learning literature. Known by practitioners that overparameterized neural networks are easy to learn, in the past few years there have been important theoretical developments in the analysis of overparameterized neural networks. In particular, it was shown that such systems behave like convex systems under various restricted settings, such as for two-layer NNs, and when learning is restricted locally in the so-called neural tangent kernel space around specialized initializations. This paper discusses some of these recent progresses leading to significant better understanding of neural networks. We will focus on the analysis of two-layer neural networks, and explain the key mathematical models, with their algorithmic implications. We will then discuss challenges in understanding deep neural networks and some current research directions. |
Automated cross-sectional view selection in CT angiography of aortic dissections with uncertainty awareness and retrospective clinical annotations | Objective: Surveillance imaging of chronic aortic diseases, such as dissections, relies on obtaining and comparing cross-sectional diameter measurements at predefined aortic landmarks, over time. Due to a lack of robust tools, the orientation of the cross-sectional planes is defined manually by highly trained operators. We show how manual annotations routinely collected in a clinic can be efficiently used to ease this task, despite the presence of a non-negligible interoperator variability in the measurements. Impact: Ill-posed but repetitive imaging tasks can be eased or automated by leveraging imperfect, retrospective clinical annotations. Methodology: In this work, we combine convolutional neural networks and uncertainty quantification methods to predict the orientation of such cross-sectional planes. We use clinical data randomly processed by 11 operators for training, and test on a smaller set processed by 3 independent operators to assess interoperator variability. Results: Our analysis shows that manual selection of cross-sectional planes is characterized by 95% limits of agreement (LOA) of $10.6^\circ$ and $21.4^\circ$ per angle. Our method showed to decrease static error by $3.57^\circ$ ($40.2$%) and $4.11^\circ$ ($32.8$%) against state of the art and LOA by $5.4^\circ$ ($49.0$%) and $16.0^\circ$ ($74.6$%) against manual processing. Conclusion: This suggests that pre-existing annotations can be an inexpensive resource in clinics to ease ill-posed and repetitive tasks like cross-section extraction for surveillance of aortic dissections. |
A Survey on Distributed Machine Learning | The demand for artificial intelligence has grown significantly over the last decade and this growth has been fueled by advances in machine learning techniques and the ability to leverage hardware acceleration. However, in order to increase the quality of predictions and render machine learning solutions feasible for more complex applications, a substantial amount of training data is required. Although small machine learning models can be trained with modest amounts of data, the input for training larger models such as neural networks grows exponentially with the number of parameters. Since the demand for processing training data has outpaced the increase in computation power of computing machinery, there is a need for distributing the machine learning workload across multiple machines, and turning the centralized into a distributed system. These distributed systems present new challenges, first and foremost the efficient parallelization of the training process and the creation of a coherent model. This article provides an extensive overview of the current state-of-the-art in the field by outlining the challenges and opportunities of distributed machine learning over conventional (centralized) machine learning, discussing the techniques used for distributed machine learning, and providing an overview of the systems that are available. |
Analytical bounds on the local Lipschitz constants of affine-ReLU functions | In this paper, we determine analytical bounds on the local Lipschitz constants of of affine functions composed with rectified linear units (ReLUs). Affine-ReLU functions represent a widely used layer in deep neural networks, due to the fact that convolution, fully-connected, and normalization functions are all affine, and are often followed by a ReLU activation function. Using an analytical approach, we mathematically determine upper bounds on the local Lipschitz constant of an affine-ReLU function, show how these bounds can be combined to determine a bound on an entire network, and discuss how the bounds can be efficiently computed, even for larger layers and networks. We show several examples by applying our results to AlexNet, as well as several smaller networks based on the MNIST and CIFAR-10 datasets. The results show that our method produces tighter bounds than the standard conservative bound (i.e. the product of the spectral norms of the layers' linear matrices), especially for small perturbations. |
An Efficient Technique for Image Captioning using Deep Neural Network | With the huge expansion of internet and trillions of gigabytes of data generated every single day, the needs for the development of various tools has become mandatory in order to maintain system adaptability to rapid changes. One of these tools is known as Image Captioning. Every entity in internet must be properly identified and managed and therefore in the case of image data, automatic captioning for identification is required. Similarly, content generation for missing labels, image classification and artificial languages all requires the process of Image Captioning. This paper discusses an efficient and unique way to perform automatic image captioning on individual image and discusses strategies to improve its performances and functionalities. |
Instance-Level Microtubule Tracking | We propose a new method of instance-level microtubule (MT) tracking in time-lapse image series using recurrent attention. Our novel deep learning algorithm segments individual MTs at each frame. Segmentation results from successive frames are used to assign correspondences among MTs. This ultimately generates a distinct path trajectory for each MT through the frames. Based on these trajectories, we estimate MT velocities. To validate our proposed technique, we conduct experiments using real and simulated data. We use statistics derived from real time-lapse series of MT gliding assays to simulate realistic MT time-lapse image series in our simulated data. This dataset is employed as pre-training and hyperparameter optimization for our network before training on the real data. Our experimental results show that the proposed supervised learning algorithm improves the precision for MT instance velocity estimation drastically to 71.3% from the baseline result (29.3%). We also demonstrate how the inclusion of temporal information into our deep network can reduce the false negative rates from 67.8% (baseline) down to 28.7% (proposed). Our findings in this work are expected to help biologists characterize the spatial arrangement of MTs, specifically the effects of MT-MT interactions. |
CHALLENGER: Training with Attribution Maps | We show that utilizing attribution maps for training neural networks can improve regularization of models and thus increase performance. Regularization is key in deep learning, especially when training complex models on relatively small datasets. In order to understand inner workings of neural networks, attribution methods such as Layer-wise Relevance Propagation (LRP) have been extensively studied, particularly for interpreting the relevance of input features. We introduce Challenger, a module that leverages the explainable power of attribution maps in order to manipulate particularly relevant input patterns. Therefore, exposing and subsequently resolving regions of ambiguity towards separating classes on the ground-truth data manifold, an issue that arises particularly when training models on rather small datasets. Our Challenger module increases model performance through building more diverse filters within the network and can be applied to any input data domain. We demonstrate that our approach results in substantially better classification as well as calibration performance on datasets with only a few samples up to datasets with thousands of samples. In particular, we show that our generic domain-independent approach yields state-of-the-art results in vision, natural language processing and on time series tasks. |
Intelligent Arxiv: Sort daily papers by learning users topics preference | Current daily paper releases are becoming increasingly large and areas of research are growing in diversity. This makes it harder for scientists to keep up to date with current state of the art and identify relevant work within their lines of interest. The goal of this article is to address this problem using Machine Learning techniques. We model a scientific paper to be built as a combination of different scientific knowledge from diverse topics into a new problem. In light of this, we implement the unsupervised Machine Learning technique of Latent Dirichlet Allocation (LDA) on the corpus of papers in a given field to: i) define and extract underlying topics in the corpus; ii) get the topics weight vector for each paper in the corpus; and iii) get the topics weight vector for new papers. By registering papers preferred by a user, we build a user vector of weights using the information of the vectors of the selected papers. Hence, by performing an inner product between the user vector and each paper in the daily Arxiv release, we can sort the papers according to the user preference on the underlying topics. We have created the website IArxiv.org where users can read sorted daily Arxiv releases (and more) while the algorithm learns each users preference, yielding a more accurate sorting every day. Current IArxiv.org version runs on Arxiv categories astro-ph, gr-qc, hep-ph and hep-th and we plan to extend to others. We propose several new useful and relevant implementations to be additionally developed as well as new Machine Learning techniques beyond LDA to further improve the accuracy of this new tool. |
CTC Variations Through New WFST Topologies | This paper presents novel Weighted Finite-State Transducer (WFST) topologies to implement Connectionist Temporal Classification (CTC)-like algorithms for automatic speech recognition. Three new CTC variants are proposed: (1) the "compact-CTC", in which direct transitions between units are replaced with <epsilon> back-off transitions; (2) the "minimal-CTC", that only adds <blank> self-loops when used in WFST-composition; and (3) the "selfless-CTC" variants, which disallows self-loop for non-blank units. Compact-CTC allows for 1.5 times smaller WFST decoding graphs and reduces memory consumption by two times when training CTC models with the LF-MMI objective without hurting the recognition accuracy. Minimal-CTC reduces graph size and memory consumption by two and four times for the cost of a small accuracy drop. Using selfless-CTC can improve the accuracy for wide context window models. |
Deep Spatio-Temporal Random Fields for Efficient Video Segmentation | In this work we introduce a time- and memory-efficient method for structured prediction that couples neuron decisions across both space at time. We show that we are able to perform exact and efficient inference on a densely connected spatio-temporal graph by capitalizing on recent advances on deep Gaussian Conditional Random Fields (GCRFs). Our method, called VideoGCRF is (a) efficient, (b) has a unique global minimum, and (c) can be trained end-to-end alongside contemporary deep networks for video understanding. We experiment with multiple connectivity patterns in the temporal domain, and present empirical improvements over strong baselines on the tasks of both semantic and instance segmentation of videos. |
Decoy Selection for Protein Structure Prediction Via Extreme Gradient Boosting and Ranking | Identifying one or more biologically-active/native decoys from millions of non-native decoys is one of the major challenges in computational structural biology. The extreme lack of balance in positive and negative samples (native and non-native decoys) in a decoy set makes the problem even more complicated. Consensus methods show varied success in handling the challenge of decoy selection despite some issues associated with clustering large decoy sets and decoy sets that do not show much structural similarity. Recent investigations into energy landscape-based decoy selection approaches show promises. However, lack of generalization over varied test cases remains a bottleneck for these methods. We propose a novel decoy selection method, ML-Select, a machine learning framework that exploits the energy landscape associated with the structure space probed through a template-free decoy generation. The proposed method outperforms both clustering and energy ranking-based methods, all the while consistently offering better performance on varied test-cases. Moreover, ML-Select shows promising results even for the decoy sets consisting of mostly low-quality decoys. ML-Select is a useful method for decoy selection. This work suggests further research in finding more effective ways to adopt machine learning frameworks in achieving robust performance for decoy selection in template-free protein structure prediction. |
Patch-level Representation Learning for Self-supervised Vision Transformers | Recent self-supervised learning (SSL) methods have shown impressive results in learning visual representations from unlabeled images. This paper aims to improve their performance further by utilizing the architectural advantages of the underlying neural network, as the current state-of-the-art visual pretext tasks for SSL do not enjoy the benefit, i.e., they are architecture-agnostic. In particular, we focus on Vision Transformers (ViTs), which have gained much attention recently as a better architectural choice, often outperforming convolutional networks for various visual tasks. The unique characteristic of ViT is that it takes a sequence of disjoint patches from an image and processes patch-level representations internally. Inspired by this, we design a simple yet effective visual pretext task, coined SelfPatch, for learning better patch-level representations. To be specific, we enforce invariance against each patch and its neighbors, i.e., each patch treats similar neighboring patches as positive samples. Consequently, training ViTs with SelfPatch learns more semantically meaningful relations among patches (without using human-annotated labels), which can be beneficial, in particular, to downstream tasks of a dense prediction type. Despite its simplicity, we demonstrate that it can significantly improve the performance of existing SSL methods for various visual tasks, including object detection and semantic segmentation. Specifically, SelfPatch significantly improves the recent self-supervised ViT, DINO, by achieving +1.3 AP on COCO object detection, +1.2 AP on COCO instance segmentation, and +2.9 mIoU on ADE20K semantic segmentation. |
The Information Geometry of Mirror Descent | Information geometry applies concepts in differential geometry to probability and statistics and is especially useful for parameter estimation in exponential families where parameters are known to lie on a Riemannian manifold. Connections between the geometric properties of the induced manifold and statistical properties of the estimation problem are well-established. However developing first-order methods that scale to larger problems has been less of a focus in the information geometry community. The best known algorithm that incorporates manifold structure is the second-order natural gradient descent algorithm introduced by Amari. On the other hand, stochastic approximation methods have led to the development of first-order methods for optimizing noisy objective functions. A recent generalization of the Robbins-Monro algorithm known as mirror descent, developed by Nemirovski and Yudin is a first order method that induces non-Euclidean geometries. However current analysis of mirror descent does not precisely characterize the induced non-Euclidean geometry nor does it consider performance in terms of statistical relative efficiency. In this paper, we prove that mirror descent induced by Bregman divergences is equivalent to the natural gradient descent algorithm on the dual Riemannian manifold. Using this equivalence, it follows that (1) mirror descent is the steepest descent direction along the Riemannian manifold of the exponential family; (2) mirror descent with log-likelihood loss applied to parameter estimation in exponential families asymptotically achieves the classical Cram\'er-Rao lower bound and (3) natural gradient descent for manifolds corresponding to exponential families can be implemented as a first-order method through mirror descent. |
Transfer Learning Using Ensemble Neural Networks for Organic Solar Cell Screening | Organic Solar Cells are a promising technology for solving the clean energy crisis in the world. However, generating candidate chemical compounds for solar cells is a time-consuming process requiring thousands of hours of laboratory analysis. For a solar cell, the most important property is the power conversion efficiency which is dependent on the highest occupied molecular orbitals (HOMO) values of the donor molecules. Recently, machine learning techniques have proved to be very useful in building predictive models for HOMO values of donor structures of Organic Photovoltaic Cells (OPVs). Since experimental datasets are limited in size, current machine learning models are trained on data derived from calculations based on density functional theory (DFT). Molecular line notations such as SMILES or InChI are popular input representations for describing the molecular structure of donor molecules. The two types of line representations encode different information, such as SMILES defines the bond types while InChi defines protonation. In this work, we present an ensemble deep neural network architecture, called SINet, which harnesses both the SMILES and InChI molecular representations to predict HOMO values and leverage the potential of transfer learning from a sizeable DFT-computed dataset- Harvard CEP to build more robust predictive models for relatively smaller HOPV datasets. Harvard CEP dataset contains molecular structures and properties for 2.3 million candidate donor structures for OPV while HOPV contains DFT-computed and experimental values of 350 and 243 molecules respectively. Our results demonstrate significant performance improvement from the use of transfer learning and leveraging both molecular representations. |
Blockchain-Federated-Learning and Deep Learning Models for COVID-19 detection using CT Imaging | With the increase of COVID-19 cases worldwide, an effective way is required to diagnose COVID-19 patients. The primary problem in diagnosing COVID-19 patients is the shortage and reliability of testing kits, due to the quick spread of the virus, medical practitioners are facing difficulty identifying the positive cases. The second real-world problem is to share the data among the hospitals globally while keeping in view the privacy concerns of the organizations. Building a collaborative model and preserving privacy are major concerns for training a global deep learning model. This paper proposes a framework that collects a small amount of data from different sources (various hospitals) and trains a global deep learning model using blockchain based federated learning. Blockchain technology authenticates the data and federated learning trains the model globally while preserving the privacy of the organization. First, we propose a data normalization technique that deals with the heterogeneity of data as the data is gathered from different hospitals having different kinds of CT scanners. Secondly, we use Capsule Network-based segmentation and classification to detect COVID-19 patients. Thirdly, we design a method that can collaboratively train a global model using blockchain technology with federated learning while preserving privacy. Additionally, we collected real-life COVID-19 patients data, which is, open to the research community. The proposed framework can utilize up-to-date data which improves the recognition of computed tomography (CT) images. Finally, our results demonstrate a better performance to detect COVID-19 patients. |
Sharp threshold for alignment of graph databases with Gaussian weights | We study the fundamental limits for reconstruction in weighted graph (or matrix) database alignment. We consider a model of two graphs where $\pi^*$ is a planted uniform permutation and all pairs of edge weights $(A_{i,j}, B_{\pi^*(i),\pi^*(j)})_{1 \leq i<j \leq n}$ are i.i.d. pairs of Gaussian variables with zero mean, unit variance and correlation parameter $\rho \in [0,1]$. We prove that there is a sharp threshold for exact recovery of $\pi^*$: if $n \rho^2 \geq (4+\epsilon) \log n + \omega(1)$ for some $\epsilon>0$, there is an estimator $\hat{\pi}$ -- namely the MAP estimator -- based on the observation of databases $A,B$ that achieves exact reconstruction with high probability. Conversely, if $n \rho^2 \leq 4 \log n - \log \log n - \omega(1)$, then any estimator $\hat{\pi}$ verifies $\hat{\pi}=\pi$ with probability $o(1)$. This result shows that the information-theoretic threshold for exact recovery is the same as the one obtained for detection in a recent work by Wu et al. (2020): in other words, for Gaussian weighted graph alignment, the problem of reconstruction is not more difficult than that of detection. Though the reconstruction task was already well understood for vector-shaped database alignment (that is taking signal of the form $(u_i, v_{\pi^*(i)})_{1 \leq i\leq n}$ where $(u_i, v_{\pi^*(i)})$ are i.i.d. pairs in $\mathbb{R}^{d_u} \times \mathbb{R}^{d_v}$), its formulation for graph (or matrix) databases brings a drastically different problem for which the hard phase is conjectured to be wide. The proofs build upon the analysis of the MAP estimator and the second moment method, together with the study of the correlation structure of energies of permutations. |
Affirmative Algorithms: The Legal Grounds for Fairness as Awareness | While there has been a flurry of research in algorithmic fairness, what is less recognized is that modern antidiscrimination law may prohibit the adoption of such techniques. We make three contributions. First, we discuss how such approaches will likely be deemed "algorithmic affirmative action," posing serious legal risks of violating equal protection, particularly under the higher education jurisprudence. Such cases have increasingly turned toward anticlassification, demanding "individualized consideration" and barring formal, quantitative weights for race regardless of purpose. This case law is hence fundamentally incompatible with fairness in machine learning. Second, we argue that the government-contracting cases offer an alternative grounding for algorithmic fairness, as these cases permit explicit and quantitative race-based remedies based on historical discrimination by the actor. Third, while limited, this doctrinal approach also guides the future of algorithmic fairness, mandating that adjustments be calibrated to the entity's responsibility for historical discrimination causing present-day disparities. The contractor cases provide a legally viable path for algorithmic fairness under current constitutional doctrine but call for more research at the intersection of algorithmic fairness and causal inference to ensure that bias mitigation is tailored to specific causes and mechanisms of bias. |
Adversarial Audio Synthesis with Complex-valued Polynomial Networks | Time-frequency (TF) representations in audio synthesis have been increasingly modeled with real-valued networks. However, overlooking the complex-valued nature of TF representations can result in suboptimal performance and require additional modules (e.g., for modeling the phase). To this end, we introduce complex-valued polynomial networks, called APOLLO, that integrate such complex-valued representations in a natural way. Concretely, APOLLO captures high-order correlations of the input elements using high-order tensors as scaling parameters. By leveraging standard tensor decompositions, we derive different architectures and enable modeling richer correlations. We outline such architectures and showcase their performance in audio generation across four benchmarks. As a highlight, APOLLO results in $17.5\%$ improvement over adversarial methods and $8.2\%$ over the state-of-the-art diffusion models on SC09 dataset in audio generation. Our models can encourage the systematic design of other efficient architectures on the complex field. |
Deep Learning Acceleration Techniques for Real Time Mobile Vision Applications | Deep Learning (DL) has become a crucial technology for Artificial Intelligence (AI). It is a powerful technique to automatically extract high-level features from complex data which can be exploited for applications such as computer vision, natural language processing, cybersecurity, communications, and so on. For the particular case of computer vision, several algorithms like object detection in real time videos have been proposed and they work well on Desktop GPUs and distributed computing platforms. However these algorithms are still heavy for mobile and embedded visual applications. The rapid spreading of smart portable devices and the emerging 5G network are introducing new smart multimedia applications in mobile environments. As a consequence, the possibility of implementing deep neural networks to mobile environments has attracted a lot of researchers. This paper presents emerging deep learning acceleration techniques that can enable the delivery of real time visual recognition into the hands of end users, anytime and anywhere. |
BIMCV COVID-19+: a large annotated dataset of RX and CT images from COVID-19 patients | This paper describes BIMCV COVID-19+, a large dataset from the Valencian Region Medical ImageBank (BIMCV) containing chest X-ray images CXR (CR, DX) and computed tomography (CT) imaging of COVID-19+ patients along with their radiological findings and locations, pathologies, radiological reports (in Spanish), DICOM metadata, Polymerase chain reaction (PCR), Immunoglobulin G (IgG) and Immunoglobulin M (IgM) diagnostic antibody tests. The findings have been mapped onto standard Unified Medical Language System (UMLS) terminology and cover a wide spectrum of thoracic entities, unlike the considerably more reduced number of entities annotated in previous datasets. Images are stored in high resolution and entities are localized with anatomical labels and stored in a Medical Imaging Data Structure (MIDS) format. In addition, 10 images were annotated by a team of radiologists to include semantic segmentation of radiological findings. This first iteration of the database includes 1,380 CX, 885 DX and 163 CT studies from 1,311 COVID-19+ patients. This is, to the best of our knowledge, the largest COVID-19+ dataset of images available in an open format. The dataset can be downloaded from http://bimcv.cipf.es/bimcv-projects/bimcv-covid19. |
Robust Subspace System Identification via Weighted Nuclear Norm Optimization | Subspace identification is a classical and very well studied problem in system identification. The problem was recently posed as a convex optimization problem via the nuclear norm relaxation. Inspired by robust PCA, we extend this framework to handle outliers. The proposed framework takes the form of a convex optimization problem with an objective that trades off fit, rank and sparsity. As in robust PCA, it can be problematic to find a suitable regularization parameter. We show how the space in which a suitable parameter should be sought can be limited to a bounded open set of the two dimensional parameter space. In practice, this is very useful since it restricts the parameter space that is needed to be surveyed. |
MDP Homomorphic Networks: Group Symmetries in Reinforcement Learning | This paper introduces MDP homomorphic networks for deep reinforcement learning. MDP homomorphic networks are neural networks that are equivariant under symmetries in the joint state-action space of an MDP. Current approaches to deep reinforcement learning do not usually exploit knowledge about such structure. By building this prior knowledge into policy and value networks using an equivariance constraint, we can reduce the size of the solution space. We specifically focus on group-structured symmetries (invertible transformations). Additionally, we introduce an easy method for constructing equivariant network layers numerically, so the system designer need not solve the constraints by hand, as is typically done. We construct MDP homomorphic MLPs and CNNs that are equivariant under either a group of reflections or rotations. We show that such networks converge faster than unstructured baselines on CartPole, a grid world and Pong. |
A Distributed Optimisation Framework Combining Natural Gradient with Hessian-Free for Discriminative Sequence Training | This paper presents a novel natural gradient and Hessian-free (NGHF) optimisation framework for neural network training that can operate efficiently in a distributed manner. It relies on the linear conjugate gradient (CG) algorithm to combine the natural gradient (NG) method with local curvature information from Hessian-free (HF) or other second-order methods. A solution to a numerical issue in CG allows effective parameter updates to be generated with far fewer CG iterations than usually used (e.g. 5-8 instead of 200). This work also presents a novel preconditioning approach to improve the progress made by individual CG iterations for models with shared parameters. Although applicable to other training losses and model structures, NGHF is investigated in this paper for lattice-based discriminative sequence training for hybrid hidden Markov model acoustic models using a standard recurrent neural network, long short-term memory, and time delay neural network models for output probability calculation. Automatic speech recognition experiments are reported on the multi-genre broadcast data set for a range of different acoustic model types. These experiments show that NGHF achieves larger word error rate reductions than standard stochastic gradient descent or Adam, while requiring orders of magnitude fewer parameter updates. |
TResNet: High Performance GPU-Dedicated Architecture | Many deep learning models, developed in recent years, reach higher ImageNet accuracy than ResNet50, with fewer or comparable FLOPS count. While FLOPs are often seen as a proxy for network efficiency, when measuring actual GPU training and inference throughput, vanilla ResNet50 is usually significantly faster than its recent competitors, offering better throughput-accuracy trade-off. In this work, we introduce a series of architecture modifications that aim to boost neural networks' accuracy, while retaining their GPU training and inference efficiency. We first demonstrate and discuss the bottlenecks induced by FLOPs-optimizations. We then suggest alternative designs that better utilize GPU structure and assets. Finally, we introduce a new family of GPU-dedicated models, called TResNet, which achieve better accuracy and efficiency than previous ConvNets. Using a TResNet model, with similar GPU throughput to ResNet50, we reach 80.8 top-1 accuracy on ImageNet. Our TResNet models also transfer well and achieve state-of-the-art accuracy on competitive single-label classification datasets such as Stanford cars (96.0%), CIFAR-10 (99.0%), CIFAR-100 (91.5%) and Oxford-Flowers (99.1%). They also perform well on multi-label classification and object detection tasks. Implementation is available at: https://github.com/mrT23/TResNet. |
Visual Analytics and Human Involvement in Machine Learning | The rapidly developing AI systems and applications still require human involvement in practically all parts of the analytics process. Human decisions are largely based on visualizations, providing data scientists details of data properties and the results of analytical procedures. Different visualizations are used in the different steps of the Machine Learning (ML) process. The decision which visualization to use depends on factors, such as the data domain, the data model and the step in the ML process. In this chapter, we describe the seven steps in the ML process and review different visualization techniques that are relevant for the different steps for different types of data, models and purposes. |
Learning Large-scale Subsurface Simulations with a Hybrid Graph Network Simulator | Subsurface simulations use computational models to predict the flow of fluids (e.g., oil, water, gas) through porous media. These simulations are pivotal in industrial applications such as petroleum production, where fast and accurate models are needed for high-stake decision making, for example, for well placement optimization and field development planning. Classical finite difference numerical simulators require massive computational resources to model large-scale real-world reservoirs. Alternatively, streamline simulators and data-driven surrogate models are computationally more efficient by relying on approximate physics models, however they are insufficient to model complex reservoir dynamics at scale. Here we introduce Hybrid Graph Network Simulator (HGNS), which is a data-driven surrogate model for learning reservoir simulations of 3D subsurface fluid flows. To model complex reservoir dynamics at both local and global scale, HGNS consists of a subsurface graph neural network (SGNN) to model the evolution of fluid flows, and a 3D-U-Net to model the evolution of pressure. HGNS is able to scale to grids with millions of cells per time step, two orders of magnitude higher than previous surrogate models, and can accurately predict the fluid flow for tens of time steps (years into the future). Using an industry-standard subsurface flow dataset (SPE-10) with 1.1 million cells, we demonstrate that HGNS is able to reduce the inference time up to 18 times compared to standard subsurface simulators, and that it outperforms other learning-based models by reducing long-term prediction errors by up to 21%. |
Low-rank Interaction with Sparse Additive Effects Model for Large Data Frames | Many applications of machine learning involve the analysis of large data frames-matrices collecting heterogeneous measurements (binary, numerical, counts, etc.) across samples-with missing values. Low-rank models, as studied by Udell et al. [30], are popular in this framework for tasks such as visualization, clustering and missing value imputation. Yet, available methods with statistical guarantees and efficient optimization do not allow explicit modeling of main additive effects such as row and column, or covariate effects. In this paper, we introduce a low-rank interaction and sparse additive effects (LORIS) model which combines matrix regression on a dictionary and low-rank design, to estimate main effects and interactions simultaneously. We provide statistical guarantees in the form of upper bounds on the estimation error of both components. Then, we introduce a mixed coordinate gradient descent (MCGD) method which provably converges sub-linearly to an optimal solution and is computationally efficient for large scale data sets. We show on simulated and survey data that the method has a clear advantage over current practices, which consist in dealing separately with additive effects in a preprocessing step. |
Fully Dense Neural Network for the Automatic Modulation Recognition | Nowadays, we mainly use various convolution neural network (CNN) structures to extract features from radio data or spectrogram in AMR. Based on expert experience and spectrograms, they not only increase the difficulty of preprocessing, but also consume a lot of memory. In order to directly use in-phase and quadrature (IQ) data obtained by the receiver and enhance the efficiency of network extraction features to improve the recognition rate of modulation mode, this paper proposes a new network structure called Fully Dense Neural Network (FDNN). This network uses residual blocks to extract features, dense connect to reduce model size, and adds attentions mechanism to recalibrate. Experiments on RML2016.10a show that this network has a higher recognition rate and lower model complexity. And it shows that the FDNN model with dense connections can not only extract features effectively but also greatly reduce model parameters, which also provides a significant contribution for the application of deep learning to the intelligent radio system. |
Identification of complex mixtures for Raman spectroscopy using a novel scheme based on a new multi-label deep neural network | With noisy environment caused by fluoresence and additive white noise as well as complicated spectrum fingerprints, the identification of complex mixture materials remains a major challenge in Raman spectroscopy application. In this paper, we propose a new scheme based on a constant wavelet transform (CWT) and a deep network for classifying complex mixture. The scheme first transforms the noisy Raman spectrum to a two-dimensional scale map using CWT. A multi-label deep neural network model (MDNN) is then applied for classifying material. The proposed model accelerates the feature extraction and expands the feature graph using the global averaging pooling layer. The Sigmoid function is implemented in the last layer of the model. The MDNN model was trained, validated and tested with data collected from the samples prepared from substances in palm oil. During training and validating process, data augmentation is applied to overcome the imbalance of data and enrich the diversity of Raman spectra. From the test results, it is found that the MDNN model outperforms previously proposed deep neural network models in terms of Hamming loss, one error, coverage, ranking loss, average precision, F1 macro averaging and F1 micro averaging, respectively. The average detection time obtained from our model is 5.31 s, which is much faster than the detection time of the previously proposed models. |
KDLSQ-BERT: A Quantized Bert Combining Knowledge Distillation with Learned Step Size Quantization | Recently, transformer-based language models such as BERT have shown tremendous performance improvement for a range of natural language processing tasks. However, these language models usually are computation expensive and memory intensive during inference. As a result, it is difficult to deploy them on resource-restricted devices. To improve the inference performance, as well as reduce the model size while maintaining the model accuracy, we propose a novel quantization method named KDLSQ-BERT that combines knowledge distillation (KD) with learned step size quantization (LSQ) for language model quantization. The main idea of our method is that the KD technique is leveraged to transfer the knowledge from a "teacher" model to a "student" model when exploiting LSQ to quantize that "student" model during the quantization training process. Extensive experiment results on GLUE benchmark and SQuAD demonstrate that our proposed KDLSQ-BERT not only performs effectively when doing different bit (e.g. 2-bit $\sim$ 8-bit) quantization, but also outperforms the existing BERT quantization methods, and even achieves comparable performance as the full-precision base-line model while obtaining 14.9x compression ratio. Our code will be public available. |
GANterfactual - Counterfactual Explanations for Medical Non-Experts using Generative Adversarial Learning | With the ongoing rise of machine learning, the need for methods for explaining decisions made by artificial intelligence systems is becoming a more and more important topic. Especially for image classification tasks, many state-of-the-art tools to explain such classifiers rely on visual highlighting of important areas of the input data. Contrary, counterfactual explanation systems try to enable a counterfactual reasoning by modifying the input image in a way such that the classifier would have made a different prediction. By doing so, the users of counterfactual explanation systems are equipped with a completely different kind of explanatory information. However, methods for generating realistic counterfactual explanations for image classifiers are still rare. Especially in medical contexts, where relevant information often consists of textural and structural information, high-quality counterfactual images have the potential to give meaningful insights into decision processes. In this work, we present GANterfactual, an approach to generate such counterfactual image explanations based on adversarial image-to-image translation techniques. Additionally, we conduct a user study to evaluate our approach in an exemplary medical use case. Our results show that, in the chosen medical use-case, counterfactual explanations lead to significantly better results regarding mental models, explanation satisfaction, trust, emotions, and self-efficacy than two state-of-the-art systems that work with saliency maps, namely LIME and LRP. |
Pattern Matching for Self- Tuning of MapReduce Jobs | In this paper, we study CPU utilization time patterns of several MapReduce applications. After extracting running patterns of several applications, they are saved in a reference database to be later used to tweak system parameters to efficiently execute unknown applications in future. To achieve this goal, CPU utilization patterns of new applications are compared with the already known ones in the reference database to find/predict their most probable execution patterns. Because of different patterns lengths, the Dynamic Time Warping (DTW) is utilized for such comparison; a correlation analysis is then applied to DTWs outcomes to produce feasible similarity patterns. Three real applications (WordCount, Exim Mainlog parsing and Terasort) are used to evaluate our hypothesis in tweaking system parameters in executing similar applications. Results were very promising and showed effectiveness of our approach on pseudo-distributed MapReduce platforms. |
Knowledge Federation: A Unified and Hierarchical Privacy-Preserving AI Framework | With strict protections and regulations of data privacy and security, conventional machine learning based on centralized datasets is confronted with significant challenges, making artificial intelligence (AI) impractical in many mission-critical and data-sensitive scenarios, such as finance, government, and health. In the meantime, tremendous datasets are scattered in isolated silos in various industries, organizations, different units of an organization, or different branches of an international organization. These valuable data resources are well underused. To advance AI theories and applications, we propose a comprehensive framework (called Knowledge Federation - KF) to address these challenges by enabling AI while preserving data privacy and ownership. Beyond the concepts of federated learning and secure multi-party computation, KF consists of four levels of federation: (1) information level, low-level statistics and computation of data, meeting the requirements of simple queries, searching and simplistic operators; (2) model level, supporting training, learning, and inference; (3) cognition level, enabling abstract feature representation at various levels of abstractions and contexts; (4) knowledge level, fusing knowledge discovery, representation, and reasoning. We further clarify the relationship and differentiation between knowledge federation and other related research areas. We have developed a reference implementation of KF, called iBond Platform, to offer a production-quality KF platform to enable industrial applications in finance, insurance et al. The iBond platform will also help establish the KF community and a comprehensive ecosystem and usher in a novel paradigm shift towards secure, privacy-preserving and responsible AI. As far as we know, knowledge federation is the first hierarchical and unified framework for secure multi-party computing and learning. |
Applying Artificial Intelligence for Age Estimation in Digital Forensic Investigations | The precise age estimation of child sexual abuse and exploitation (CSAE) victims is one of the most significant digital forensic challenges. Investigators often need to determine the age of victims by looking at images and interpreting the sexual development stages and other human characteristics. The main priority - safeguarding children -- is often negatively impacted by a huge forensic backlog, cognitive bias and the immense psychological stress that this work can entail. This paper evaluates existing facial image datasets and proposes a new dataset tailored to the needs of similar digital forensic research contributions. This small, diverse dataset of 0 to 20-year-old individuals contains 245 images and is merged with 82 unique images from the FG-NET dataset, thus achieving a total of 327 images with high image diversity and low age range density. The new dataset is tested on the Deep EXpectation (DEX) algorithm pre-trained on the IMDB-WIKI dataset. The overall results for young adolescents aged 10 to 15 and older adolescents/adults aged 16 to 20 are very encouraging -- achieving MAEs as low as 1.79, but also suggest that the accuracy for children aged 0 to 10 needs further work. In order to determine the efficacy of the prototype, valuable input of four digital forensic experts, including two forensic investigators, has been taken into account to improve age estimation results. Further research is required to extend datasets both concerning image density and the equal distribution of factors such as gender and racial diversity. |
Wave-based extreme deep learning based on non-linear time-Floquet entanglement | Wave-based analog signal processing holds the promise of extremely fast, on-the-fly, power-efficient data processing, occurring as a wave propagates through an artificially engineered medium. Yet, due to the fundamentally weak non-linearities of traditional wave materials, such analog processors have been so far largely confined to simple linear projections such as image edge detection or matrix multiplications. Complex neuromorphic computing tasks, which inherently require strong non-linearities, have so far remained out-of-reach of wave-based solutions, with a few attempts that implemented non-linearities on the digital front, or used weak and inflexible non-linear sensors, restraining the learning performance. Here, we tackle this issue by demonstrating the relevance of Time-Floquet physics to induce a strong non-linear entanglement between signal inputs at different frequencies, enabling a power-efficient and versatile wave platform for analog extreme deep learning involving a single, uniformly modulated dielectric layer and a scattering medium. We prove the efficiency of the method for extreme learning machines and reservoir computing to solve a range of challenging learning tasks, from forecasting chaotic time series to the simultaneous classification of distinct datasets. Our results open the way for wave-based machine learning with high energy efficiency, speed, and scalability. |
Enhanced Classification Accuracy for Cardiotocogram Data with Ensemble Feature Selection and Classifier Ensemble | In this paper ensemble learning based feature selection and classifier ensemble model is proposed to improve classification accuracy. The hypothesis is that good feature sets contain features that are highly correlated with the class from ensemble feature selection to SVM ensembles which can be achieved on the performance of classification accuracy. The proposed approach consists of two phases: (i) to select feature sets that are likely to be the support vectors by applying ensemble based feature selection methods; and (ii) to construct an SVM ensemble using the selected features. The proposed approach was evaluated by experiments on Cardiotocography dataset. Four feature selection techniques were used: (i) Correlation-based, (ii) Consistency-based, (iii) ReliefF and (iv) Information Gain. Experimental results showed that using the ensemble of Information Gain feature selection and Correlation-based feature selection with SVM ensembles achieved higher classification accuracy than both single SVM classifier and ensemble feature selection with SVM classifier. |
Calibrated Top-1 Uncertainty estimates for classification by score based models | While the accuracy of modern deep learning models has significantly improved in recent years, the ability of these models to generate uncertainty estimates has not progressed to the same degree. Uncertainty methods are designed to provide an estimate of class probabilities when predicting class assignment. While there are a number of proposed methods for estimating uncertainty, they all suffer from a lack of calibration: predicted probabilities can be off from empirical ones by a few percent or more. By restricting the scope of our predictions to only the probability of Top-1 error, we can decrease the calibration error of existing methods to less than one percent. As a result, the scores of the methods also improve significantly over benchmarks. |
Precise expressions for random projections: Low-rank approximation and randomized Newton | It is often desirable to reduce the dimensionality of a large dataset by projecting it onto a low-dimensional subspace. Matrix sketching has emerged as a powerful technique for performing such dimensionality reduction very efficiently. Even though there is an extensive literature on the worst-case performance of sketching, existing guarantees are typically very different from what is observed in practice. We exploit recent developments in the spectral analysis of random matrices to develop novel techniques that provide provably accurate expressions for the expected value of random projection matrices obtained via sketching. These expressions can be used to characterize the performance of dimensionality reduction in a variety of common machine learning tasks, ranging from low-rank approximation to iterative stochastic optimization. Our results apply to several popular sketching methods, including Gaussian and Rademacher sketches, and they enable precise analysis of these methods in terms of spectral properties of the data. Empirical results show that the expressions we derive reflect the practical performance of these sketching methods, down to lower-order effects and even constant factors. |
Screening Mammogram Classification with Prior Exams | Radiologists typically compare a patient's most recent breast cancer screening exam to their previous ones in making informed diagnoses. To reflect this practice, we propose new neural network models that compare pairs of screening mammograms from the same patient. We train and evaluate our proposed models on over 665,000 pairs of images (over 166,000 pairs of exams). Our best model achieves an AUC of 0.866 in predicting malignancy in patients who underwent breast cancer screening, reducing the error rate of the corresponding baseline. |
The role of cue enhancement and frequency fine-tuning in hearing impaired phone recognition | A speech-based hearing test is designed to identify the susceptible error-prone phones for individual hearing impaired (HI) ear. Only robust tokens in the experiment noise levels had been chosen for the test. The noise-robustness of tokens is measured as SNR90 of the token, which is the signal to the speech-weighted noise ratio where a normal hearing (NH) listener would recognize the token with an accuracy of 90% on average. Two sets of tokens T1 and T2 having the same consonant-vowels but different talkers with distinct SNR90 had been presented with flat gain at listeners' most comfortable level. We studied the effects of frequency fine-tuning of the primary cue by presenting tokens of the same consonant but different vowels with similar SNR90. Additionally, we investigated the role of changing the intensity of primary cue in HI phone recognition, by presenting tokens from both sets T1 and T2. On average, 92% of tokens are improved when we replaced the CV with the same CV but with a more robust talker. Additionally, using CVs with similar SNR90, on average, tokens are improved by 75%, 71%, 63%, and 72%, when we replaced vowels /A, ae, I, E/, respectively. The confusion pattern in each case provides insight into how these changes affect the phone recognition in each HI ear. We propose to prescribe hearing aid amplification tailored to individual HI ears, based on the confusion pattern, the response from cue enhancement, and the response from frequency fine-tuning of the cue. |
Visual-Tactile Multimodality for Following Deformable Linear Objects Using Reinforcement Learning | Manipulation of deformable objects is a challenging task for a robot. It will be problematic to use a single sensory input to track the behaviour of such objects: vision can be subjected to occlusions, whereas tactile inputs cannot capture the global information that is useful for the task. In this paper, we study the problem of using vision and tactile inputs together to complete the task of following deformable linear objects, for the first time. We create a Reinforcement Learning agent using different sensing modalities and investigate how its behaviour can be boosted using visual-tactile fusion, compared to using a single sensing modality. To this end, we developed a benchmark in simulation for manipulating the deformable linear objects using multimodal sensing inputs. The policy of the agent uses distilled information, e.g., the pose of the object in both visual and tactile perspectives, instead of the raw sensing signals, so that it can be directly transferred to real environments. In this way, we disentangle the perception system and the learned control policy. Our extensive experiments show that the use of both vision and tactile inputs, together with proprioception, allows the agent to complete the task in up to 92% of cases, compared to 77% when only one of the signals is given. Our results can provide valuable insights for the future design of tactile sensors and for deformable objects manipulation. |
Online learning of both state and dynamics using ensemble Kalman filters | The reconstruction of the dynamics of an observed physical system as a surrogate model has been brought to the fore by recent advances in machine learning. To deal with partial and noisy observations in that endeavor, machine learning representations of the surrogate model can be used within a Bayesian data assimilation framework. However, these approaches require to consider long time series of observational data, meant to be assimilated all together. This paper investigates the possibility to learn both the dynamics and the state online, i.e. to update their estimates at any time, in particular when new observations are acquired. The estimation is based on the ensemble Kalman filter (EnKF) family of algorithms using a rather simple representation for the surrogate model and state augmentation. We consider the implication of learning dynamics online through (i) a global EnKF, (i) a local EnKF and (iii) an iterative EnKF and we discuss in each case issues and algorithmic solutions. We then demonstrate numerically the efficiency and assess the accuracy of these methods using one-dimensional, one-scale and two-scale chaotic Lorenz models. |
Online Multi-task Learning with Hard Constraints | We discuss multi-task online learning when a decision maker has to deal simultaneously with M tasks. The tasks are related, which is modeled by imposing that the M-tuple of actions taken by the decision maker needs to satisfy certain constraints. We give natural examples of such restrictions and then discuss a general class of tractable constraints, for which we introduce computationally efficient ways of selecting actions, essentially by reducing to an on-line shortest path problem. We briefly discuss "tracking" and "bandit" versions of the problem and extend the model in various ways, including non-additive global losses and uncountably infinite sets of tasks. |
The Tree Ensemble Layer: Differentiability meets Conditional Computation | Neural networks and tree ensembles are state-of-the-art learners, each with its unique statistical and computational advantages. We aim to combine these advantages by introducing a new layer for neural networks, composed of an ensemble of differentiable decision trees (a.k.a. soft trees). While differentiable trees demonstrate promising results in the literature, they are typically slow in training and inference as they do not support conditional computation. We mitigate this issue by introducing a new sparse activation function for sample routing, and implement true conditional computation by developing specialized forward and backward propagation algorithms that exploit sparsity. Our efficient algorithms pave the way for jointly training over deep and wide tree ensembles using first-order methods (e.g., SGD). Experiments on 23 classification datasets indicate over 10x speed-ups compared to the differentiable trees used in the literature and over 20x reduction in the number of parameters compared to gradient boosted trees, while maintaining competitive performance. Moreover, experiments on CIFAR, MNIST, and Fashion MNIST indicate that replacing dense layers in CNNs with our tree layer reduces the test loss by 7-53% and the number of parameters by 8x. We provide an open-source TensorFlow implementation with a Keras API. |
Fishr: Invariant Gradient Variances for Out-of-Distribution Generalization | Learning robust models that generalize well under changes in the data distribution is critical for real-world applications. To this end, there has been a growing surge of interest to learn simultaneously from multiple training domains - while enforcing different types of invariance across those domains. Yet, all existing approaches fail to show systematic benefits under controlled evaluation protocols. In this paper, we introduce a new regularization - named Fishr - that enforces domain invariance in the space of the gradients of the loss: specifically, the domain-level variances of gradients are matched across training domains. Our approach is based on the close relations between the gradient covariance, the Fisher Information and the Hessian of the loss: in particular, we show that Fishr eventually aligns the domain-level loss landscapes locally around the final weights. Extensive experiments demonstrate the effectiveness of Fishr for out-of-distribution generalization. Notably, Fishr improves the state of the art on the DomainBed benchmark and performs consistently better than Empirical Risk Minimization. Our code is available at https://github.com/alexrame/fishr. |
Stepwise-Refining Speech Separation Network via Fine-Grained Encoding in High-order Latent Domain | The crux of single-channel speech separation is how to encode the mixture of signals into such a latent embedding space that the signals from different speakers can be precisely separated. Existing methods for speech separation either transform the speech signals into frequency domain to perform separation or seek to learn a separable embedding space by constructing a latent domain based on convolutional filters. While the latter type of methods learning an embedding space achieves substantial improvement for speech separation, we argue that the embedding space defined by only one latent domain does not suffice to provide a thoroughly separable encoding space for speech separation. In this paper, we propose the Stepwise-Refining Speech Separation Network (SRSSN), which follows a coarse-to-fine separation framework. It first learns a 1-order latent domain to define an encoding space and thereby performs a rough separation in the coarse phase. Then the proposed SRSSN learns a new latent domain along each basis function of the existing latent domain to obtain a high-order latent domain in the refining phase, which enables our model to perform a refining separation to achieve a more precise speech separation. We demonstrate the effectiveness of our SRSSN by conducting extensive experiments, including speech separation in a clean (noise-free) setting on WSJ0-2/3mix datasets as well as in noisy/reverberant settings on WHAM!/WHAMR! datasets. Furthermore, we also perform experiments of speech recognition on separated speech signals by our model to evaluate the performance of speech separation indirectly. |
Signal Quality Assessment of Photoplethysmogram Signals using Quantum Pattern Recognition and lightweight CNN Architecture | Photoplethysmography (PPG) signal comprises physiological information related to cardiorespiratory health. However, while recording, these PPG signals are easily corrupted by motion artifacts and body movements, leading to noise enriched, poor quality signals. Therefore ensuring high-quality signals is necessary to extract cardiorespiratory information accurately. Although there exists several rule-based and Machine-Learning (ML) - based approaches for PPG signal quality estimation, those algorithms' efficacy is questionable. Thus, this work proposes a lightweight CNN architecture for signal quality assessment employing a novel Quantum pattern recognition (QPR) technique. The proposed algorithm is validated on manually annotated data obtained from the University of Queensland database. A total of 28366, 5s signal segments are preprocessed and transformed into image files of 20 x 500 pixels. The image files are treated as an input to the 2D CNN architecture. The developed model classifies the PPG signal as `good' or `bad' with an accuracy of 98.3% with 99.3% sensitivity, 94.5% specificity and 98.9% F1-score. Finally, the performance of the proposed framework is validated against the noisy `Welltory app' collected PPG database. Even in a noisy environment, the proposed architecture proved its competence. Experimental analysis concludes that a slim architecture along with a novel Spatio-temporal pattern recognition technique improve the system's performance. Hence, the proposed approach can be useful to classify good and bad PPG signals for a resource-constrained wearable implementation. |
An Approximation Algorithm for Optimal Subarchitecture Extraction | We consider the problem of finding the set of architectural parameters for a chosen deep neural network which is optimal under three metrics: parameter size, inference speed, and error rate. In this paper we state the problem formally, and present an approximation algorithm that, for a large subset of instances behaves like an FPTAS with an approximation error of $\rho \leq |{1- \epsilon}|$, and that runs in $O(|{\Xi}| + |{W^*_T}|(1 + |{\Theta}||{B}||{\Xi}|/({\epsilon\, s^{3/2})}))$ steps, where $\epsilon$ and $s$ are input parameters; $|{B}|$ is the batch size; $|{W^*_T}|$ denotes the cardinality of the largest weight set assignment; and $|{\Xi}|$ and $|{\Theta}|$ are the cardinalities of the candidate architecture and hyperparameter spaces, respectively. |
f-Domain-Adversarial Learning: Theory and Algorithms | Unsupervised domain adaptation is used in many machine learning applications where, during training, a model has access to unlabeled data in the target domain, and a related labeled dataset. In this paper, we introduce a novel and general domain-adversarial framework. Specifically, we derive a novel generalization bound for domain adaptation that exploits a new measure of discrepancy between distributions based on a variational characterization of f-divergences. It recovers the theoretical results from Ben-David et al. (2010a) as a special case and supports divergences used in practice. Based on this bound, we derive a new algorithmic framework that introduces a key correction in the original adversarial training method of Ganin et al. (2016). We show that many regularizers and ad-hoc objectives introduced over the last years in this framework are then not required to achieve performance comparable to (if not better than) state-of-the-art domain-adversarial methods. Experimental analysis conducted on real-world natural language and computer vision datasets show that our framework outperforms existing baselines, and obtains the best results for f-divergences that were not considered previously in domain-adversarial learning. |
Accelerating Deep Learning Inference via Learned Caches | Deep Neural Networks (DNNs) are witnessing increased adoption in multiple domains owing to their high accuracy in solving real-world problems. However, this high accuracy has been achieved by building deeper networks, posing a fundamental challenge to the low latency inference desired by user-facing applications. Current low latency solutions trade-off on accuracy or fail to exploit the inherent temporal locality in prediction serving workloads. We observe that caching hidden layer outputs of the DNN can introduce a form of late-binding where inference requests only consume the amount of computation needed. This enables a mechanism for achieving low latencies, coupled with an ability to exploit temporal locality. However, traditional caching approaches incur high memory overheads and lookup latencies, leading us to design learned caches - caches that consist of simple ML models that are continuously updated. We present the design of GATI, an end-to-end prediction serving system that incorporates learned caches for low-latency DNN inference. Results show that GATI can reduce inference latency by up to 7.69X on realistic workloads. |
Machine Learning on Camera Images for Fast mmWave Beamforming | Perfect alignment in chosen beam sectors at both transmit- and receive-nodes is required for beamforming in mmWave bands. Current 802.11ad WiFi and emerging 5G cellular standards spend up to several milliseconds exploring different sector combinations to identify the beam pair with the highest SNR. In this paper, we propose a machine learning (ML) approach with two sequential convolutional neural networks (CNN) that uses out-of-band information, in the form of camera images, to (i) rapidly identify the locations of the transmitter and receiver nodes, and then (ii) return the optimal beam pair. We experimentally validate this intriguing concept for indoor settings using the NI 60GHz mmwave transceiver. Our results reveal that our ML approach reduces beamforming related exploration time by 93% under different ambient lighting conditions, with an error of less than 1% compared to the time-intensive deterministic method defined by the current standards. |
A Few-shot Learning Graph Multi-Trajectory Evolution Network for Forecasting Multimodal Baby Connectivity Development from a Baseline Timepoint | Charting the baby connectome evolution trajectory during the first year after birth plays a vital role in understanding dynamic connectivity development of baby brains. Such analysis requires acquisition of longitudinal connectomic datasets. However, both neonatal and postnatal scans are rarely acquired due to various difficulties. A small body of works has focused on predicting baby brain evolution trajectory from a neonatal brain connectome derived from a single modality. Although promising, large training datasets are essential to boost model learning and to generalize to a multi-trajectory prediction from different modalities (i.e., functional and morphological connectomes). Here, we unprecedentedly explore the question: Can we design a few-shot learning-based framework for predicting brain graph trajectories across different modalities? To this aim, we propose a Graph Multi-Trajectory Evolution Network (GmTE-Net), which adopts a teacher-student paradigm where the teacher network learns on pure neonatal brain graphs and the student network learns on simulated brain graphs given a set of different timepoints. To the best of our knowledge, this is the first teacher-student architecture tailored for brain graph multi-trajectory growth prediction that is based on few-shot learning and generalized to graph neural networks (GNNs). To boost the performance of the student network, we introduce a local topology-aware distillation loss that forces the predicted graph topology of the student network to be consistent with the teacher network. Experimental results demonstrate substantial performance gains over benchmark methods. Hence, our GmTE-Net can be leveraged to predict atypical brain connectivity trajectory evolution across various modalities. Our code is available at https: //github.com/basiralab/GmTE-Net. |
The Secrets of Machine Learning: Ten Things You Wish You Had Known Earlier to be More Effective at Data Analysis | Despite the widespread usage of machine learning throughout organizations, there are some key principles that are commonly missed. In particular: 1) There are at least four main families for supervised learning: logical modeling methods, linear combination methods, case-based reasoning methods, and iterative summarization methods. 2) For many application domains, almost all machine learning methods perform similarly (with some caveats). Deep learning methods, which are the leading technique for computer vision problems, do not maintain an edge over other methods for most problems (and there are reasons why). 3) Neural networks are hard to train and weird stuff often happens when you try to train them. 4) If you don't use an interpretable model, you can make bad mistakes. 5) Explanations can be misleading and you can't trust them. 6) You can pretty much always find an accurate-yet-interpretable model, even for deep neural networks. 7) Special properties such as decision making or robustness must be built in, they don't happen on their own. 8) Causal inference is different than prediction (correlation is not causation). 9) There is a method to the madness of deep neural architectures, but not always. 10) It is a myth that artificial intelligence can do anything. |
NeuroNet: Fast and Robust Reproduction of Multiple Brain Image Segmentation Pipelines | NeuroNet is a deep convolutional neural network mimicking multiple popular and state-of-the-art brain segmentation tools including FSL, SPM, and MALPEM. The network is trained on 5,000 T1-weighted brain MRI scans from the UK Biobank Imaging Study that have been automatically segmented into brain tissue and cortical and sub-cortical structures using the standard neuroimaging pipelines. Training a single model from these complementary and partially overlapping label maps yields a new powerful "all-in-one", multi-output segmentation tool. The processing time for a single subject is reduced by an order of magnitude compared to running each individual software package. We demonstrate very good reproducibility of the original outputs while increasing robustness to variations in the input data. We believe NeuroNet could be an important tool in large-scale population imaging studies and serve as a new standard in neuroscience by reducing the risk of introducing bias when choosing a specific software package. |
Estimating and Penalizing Induced Preference Shifts in Recommender Systems | The content that a recommender system (RS) shows to users influences them. Therefore, when choosing which recommender to deploy, one is implicitly also choosing to induce specific internal states in users. Even more, systems trained via long-horizon optimization will have direct incentives to manipulate users, e.g. shift their preferences so they are easier to satisfy. In this work we focus on induced preference shifts in users. We argue that - before deployment - system designers should: estimate the shifts a recommender would induce; evaluate whether such shifts would be undesirable; and even actively optimize to avoid problematic shifts. These steps involve two challenging ingredients: estimation requires anticipating how hypothetical policies would influence user preferences if deployed - we do this by using historical user interaction data to train predictive user model which implicitly contains their preference dynamics; evaluation and optimization additionally require metrics to assess whether such influences are manipulative or otherwise unwanted - we use the notion of "safe shifts", that define a trust region within which behavior is safe. In simulated experiments, we show that our learned preference dynamics model is effective in estimating user preferences and how they would respond to new recommenders. Additionally, we show that recommenders that optimize for staying in the trust region can avoid manipulative behaviors while still generating engagement. |
Scalable Methods for 8-bit Training of Neural Networks | Quantized Neural Networks (QNNs) are often used to improve network efficiency during the inference phase, i.e. after the network has been trained. Extensive research in the field suggests many different quantization schemes. Still, the number of bits required, as well as the best quantization scheme, are yet unknown. Our theoretical analysis suggests that most of the training process is robust to substantial precision reduction, and points to only a few specific operations that require higher precision. Armed with this knowledge, we quantize the model parameters, activations and layer gradients to 8-bit, leaving at a higher precision only the final step in the computation of the weight gradients. Additionally, as QNNs require batch-normalization to be trained at high precision, we introduce Range Batch-Normalization (BN) which has significantly higher tolerance to quantization noise and improved computational complexity. Our simulations show that Range BN is equivalent to the traditional batch norm if a precise scale adjustment, which can be approximated analytically, is applied. To the best of the authors' knowledge, this work is the first to quantize the weights, activations, as well as a substantial volume of the gradients stream, in all layers (including batch normalization) to 8-bit while showing state-of-the-art results over the ImageNet-1K dataset. |
Emergent Communication Pretraining for Few-Shot Machine Translation | While state-of-the-art models that rely upon massively multilingual pretrained encoders achieve sample efficiency in downstream applications, they still require abundant amounts of unlabelled text. Nevertheless, most of the world's languages lack such resources. Hence, we investigate a more radical form of unsupervised knowledge transfer in the absence of linguistic data. In particular, for the first time we pretrain neural networks via emergent communication from referential games. Our key assumption is that grounding communication on images---as a crude approximation of real-world environments---inductively biases the model towards learning natural languages. On the one hand, we show that this substantially benefits machine translation in few-shot settings. On the other hand, this also provides an extrinsic evaluation protocol to probe the properties of emergent languages ex vitro. Intuitively, the closer they are to natural languages, the higher the gains from pretraining on them should be. For instance, in this work we measure the influence of communication success and maximum sequence length on downstream performances. Finally, we introduce a customised adapter layer and annealing strategies for the regulariser of maximum-a-posteriori inference during fine-tuning. These turn out to be crucial to facilitate knowledge transfer and prevent catastrophic forgetting. Compared to a recurrent baseline, our method yields gains of $59.0\%$$\sim$$147.6\%$ in BLEU score with only $500$ NMT training instances and $65.1\%$$\sim$$196.7\%$ with $1,000$ NMT training instances across four language pairs. These proof-of-concept results reveal the potential of emergent communication pretraining for both natural language processing tasks in resource-poor settings and extrinsic evaluation of artificial languages. |
Certified Monotonic Neural Networks | Learning monotonic models with respect to a subset of the inputs is a desirable feature to effectively address the fairness, interpretability, and generalization issues in practice. Existing methods for learning monotonic neural networks either require specifically designed model structures to ensure monotonicity, which can be too restrictive/complicated, or enforce monotonicity by adjusting the learning process, which cannot provably guarantee the learned model is monotonic on selected features. In this work, we propose to certify the monotonicity of the general piece-wise linear neural networks by solving a mixed integer linear programming problem.This provides a new general approach for learning monotonic neural networks with arbitrary model structures. Our method allows us to train neural networks with heuristic monotonicity regularizations, and we can gradually increase the regularization magnitude until the learned network is certified monotonic. Compared to prior works, our approach does not require human-designed constraints on the weight space and also yields more accurate approximation. Empirical studies on various datasets demonstrate the efficiency of our approach over the state-of-the-art methods, such as Deep Lattice Networks. |
Automated Learning Rate Scheduler for Large-batch Training | Large-batch training has been essential in leveraging large-scale datasets and models in deep learning. While it is computationally beneficial to use large batch sizes, it often requires a specially designed learning rate (LR) schedule to achieve a comparable level of performance as in smaller batch training. Especially, when the number of training epochs is constrained, the use of a large LR and a warmup strategy is critical in the final performance of large-batch training due to the reduced number of updating steps. In this work, we propose an automated LR scheduling algorithm which is effective for neural network training with a large batch size under the given epoch budget. In specific, the whole schedule consists of two phases: adaptive warmup and predefined decay, where the LR is increased until the training loss no longer decreases and decreased to zero until the end of training. Here, whether the training loss has reached the minimum value is robustly checked with Gaussian process smoothing in an online manner with a low computational burden. Coupled with adaptive stochastic optimizers such as AdamP and LAMB, the proposed scheduler successfully adjusts the LRs without cumbersome hyperparameter tuning and achieves comparable or better performances than tuned baselines on various image classification benchmarks and architectures with a wide range of batch sizes. |
Finding Efficient Swimming Strategies in a Three Dimensional Chaotic Flow by Reinforcement Learning | We apply a reinforcement learning algorithm to show how smart particles can learn approximately optimal strategies to navigate in complex flows. In this paper we consider microswimmers in a paradigmatic three-dimensional case given by a stationary superposition of two Arnold-Beltrami-Childress flows with chaotic advection along streamlines. In such a flow, we study the evolution of point-like particles which can decide in which direction to swim, while keeping the velocity amplitude constant. We show that it is sufficient to endow the swimmers with a very restricted set of actions (six fixed swimming directions in our case) to have enough freedom to find efficient strategies to move upward and escape local fluid traps. The key ingredient is the learning-from-experience structure of the algorithm, which assigns positive or negative rewards depending on whether the taken action is, or is not, profitable for the predetermined goal in the long term horizon. This is another example supporting the efficiency of the reinforcement learning approach to learn how to accomplish difficult tasks in complex fluid environments. |
The Open Vault Challenge -- Learning how to build calibration-free interactive systems by cracking the code of a vault | This demo takes the form of a challenge to the IJCAI community. A physical vault, secured by a 4-digit code, will be placed in the demo area. The author will publicly open the vault by entering the code on a touch-based interface, and as many times as requested. The challenge to the IJCAI participants will be to crack the code, open the vault, and collect its content. The interface is based on previous work on calibration-free interactive systems that enables a user to start instructing a machine without the machine knowing how to interpret the user's actions beforehand. The intent and the behavior of the human are simultaneously learned by the machine. An online demo and videos are available for readers to participate in the challenge. An additional interface using vocal commands will be revealed on the demo day, demonstrating the scalability of our approach to continuous input signals. |
State-of-Charge Estimation of a Li-Ion Battery using Deep Forward Neural Networks | This article presents two Deep Forward Networks with two and four hidden layers, respectively, that model the drive cycle of a Panasonic 18650PF lithium-ion (Li-ion) battery at a given temperature using the K-fold cross-validation method, in order to estimate the State of Charge (SOC) of the cell. The drive cycle power profile is calculated for an electric truck with a 35kWh battery pack scaled for a single 18650PF cell. We propose a machine learning workflow which is able to fight overfitting when developing deep learning models for SOC estimation. The contribution of this work is to present a methodology of building a Deep Forward Network for a lithium-ion battery and its performance assessment, which follows the best practices in machine learning. |
Data Augmentation Revisited: Rethinking the Distribution Gap between Clean and Augmented Data | Data augmentation has been widely applied as an effective methodology to improve generalization in particular when training deep neural networks. Recently, researchers proposed a few intensive data augmentation techniques, which indeed improved accuracy, yet we notice that these methods augment data have also caused a considerable gap between clean and augmented data. In this paper, we revisit this problem from an analytical perspective, for which we estimate the upper-bound of expected risk using two terms, namely, empirical risk and generalization error, respectively. We develop an understanding of data augmentation as regularization, which highlights the major features. As a result, data augmentation significantly reduces the generalization error, but meanwhile leads to a slightly higher empirical risk. On the assumption that data augmentation helps models converge to a better region, the model can benefit from a lower empirical risk achieved by a simple method, i.e., using less-augmented data to refine the model trained on fully-augmented data. Our approach achieves consistent accuracy gain on a few standard image classification benchmarks, and the gain transfers to object detection. |
Predictive Optimization with Zero-Shot Domain Adaptation | Prediction in a new domain without any training sample, called zero-shot domain adaptation (ZSDA), is an important task in domain adaptation. While prediction in a new domain has gained much attention in recent years, in this paper, we investigate another potential of ZSDA. Specifically, instead of predicting responses in a new domain, we find a description of a new domain given a prediction. The task is regarded as predictive optimization, but existing predictive optimization methods have not been extended to handling multiple domains. We propose a simple framework for predictive optimization with ZSDA and analyze the condition in which the optimization problem becomes convex optimization. We also discuss how to handle the interaction of characteristics of a domain in predictive optimization. Through numerical experiments, we demonstrate the potential usefulness of our proposed framework. |
Hierarchical Graph Representation Learning for the Prediction of Drug-Target Binding Affinity | The identification of drug-target binding affinity (DTA) has attracted increasing attention in the drug discovery process due to the more specific interpretation than binary interaction prediction. Recently, numerous deep learning-based computational methods have been proposed to predict the binding affinities between drugs and targets benefiting from their satisfactory performance. However, the previous works mainly focus on encoding biological features and chemical structures of drugs and targets, with a lack of exploiting the essential topological information from the drug-target affinity network. In this paper, we propose a novel hierarchical graph representation learning model for the drug-target binding affinity prediction, namely HGRL-DTA. The main contribution of our model is to establish a hierarchical graph learning architecture to incorporate the intrinsic properties of drug/target molecules and the topological affinities of drug-target pairs. In this architecture, we adopt a message broadcasting mechanism to integrate the hierarchical representations learned from the global-level affinity graph and the local-level molecular graph. Besides, we design a similarity-based embedding map to solve the cold start problem of inferring representations for unseen drugs and targets. Comprehensive experimental results under different scenarios indicate that HGRL-DTA significantly outperforms the state-of-the-art models and shows better model generalization among all the scenarios. |
Accountable Off-Policy Evaluation With Kernel Bellman Statistics | We consider off-policy evaluation (OPE), which evaluates the performance of a new policy from observed data collected from previous experiments, without requiring the execution of the new policy. This finds important applications in areas with high execution cost or safety concerns, such as medical diagnosis, recommendation systems and robotics. In practice, due to the limited information from off-policy data, it is highly desirable to construct rigorous confidence intervals, not just point estimation, for the policy performance. In this work, we propose a new variational framework which reduces the problem of calculating tight confidence bounds in OPE into an optimization problem on a feasible set that catches the true state-action value function with high probability. The feasible set is constructed by leveraging statistical properties of a recently proposed kernel Bellman loss (Feng et al., 2019). We design an efficient computational approach for calculating our bounds, and extend it to perform post-hoc diagnosis and correction for existing estimators. Empirical results show that our method yields tight confidence intervals in different settings. |
Control on the Manifolds of Mappings with a View to the Deep Learning | Deep learning of the Artificial Neural Networks (ANN) can be treated as a particular class of interpolation problems. The goal is to find a neural network whose input-output map approximates well the desired map on a finite or an infinite training set. Our idea consists of taking as an approximant the input-output map, which arises from a nonlinear continuous-time control system. In the limit such control system can be seen as a network with a continuum of layers, each one labelled by the time variable. The values of the controls at each instant of time are the parameters of the layer. |
Explanations of model predictions with live and breakDown packages | Complex models are commonly used in predictive modeling. In this paper we present R packages that can be used to explain predictions from complex black box models and attribute parts of these predictions to input features. We introduce two new approaches and corresponding packages for such attribution, namely live and breakDown. We also compare their results with existing implementations of state of the art solutions, namely lime that implements Locally Interpretable Model-agnostic Explanations and ShapleyR that implements Shapley values. |
Cooperative learning in multi-agent systems from intermittent measurements | Motivated by the problem of tracking a direction in a decentralized way, we consider the general problem of cooperative learning in multi-agent systems with time-varying connectivity and intermittent measurements. We propose a distributed learning protocol capable of learning an unknown vector $\mu$ from noisy measurements made independently by autonomous nodes. Our protocol is completely distributed and able to cope with the time-varying, unpredictable, and noisy nature of inter-agent communication, and intermittent noisy measurements of $\mu$. Our main result bounds the learning speed of our protocol in terms of the size and combinatorial features of the (time-varying) networks connecting the nodes. |
Multigrid Neural Architectures | We propose a multigrid extension of convolutional neural networks (CNNs). Rather than manipulating representations living on a single spatial grid, our network layers operate across scale space, on a pyramid of grids. They consume multigrid inputs and produce multigrid outputs; convolutional filters themselves have both within-scale and cross-scale extent. This aspect is distinct from simple multiscale designs, which only process the input at different scales. Viewed in terms of information flow, a multigrid network passes messages across a spatial pyramid. As a consequence, receptive field size grows exponentially with depth, facilitating rapid integration of context. Most critically, multigrid structure enables networks to learn internal attention and dynamic routing mechanisms, and use them to accomplish tasks on which modern CNNs fail. Experiments demonstrate wide-ranging performance advantages of multigrid. On CIFAR and ImageNet classification tasks, flipping from a single grid to multigrid within the standard CNN paradigm improves accuracy, while being compute and parameter efficient. Multigrid is independent of other architectural choices; we show synergy in combination with residual connections. Multigrid yields dramatic improvement on a synthetic semantic segmentation dataset. Most strikingly, relatively shallow multigrid networks can learn to directly perform spatial transformation tasks, where, in contrast, current CNNs fail. Together, our results suggest that continuous evolution of features on a multigrid pyramid is a more powerful alternative to existing CNN designs on a flat grid. |
Structured Pruning of Large Language Models | Large language models have recently achieved state of the art performance across a wide variety of natural language tasks. Meanwhile, the size of these models and their latency have significantly increased, which makes their usage costly, and raises an interesting question: do language models need to be large? We study this question through the lens of model compression. We present a generic, structured pruning approach by parameterizing each weight matrix using its low-rank factorization, and adaptively removing rank-1 components during training. On language modeling tasks, our structured approach outperforms other unstructured and block-structured pruning baselines at various compression levels, while achieving significant speedups during both training and inference. We also demonstrate that our method can be applied to pruning adaptive word embeddings in large language models, and to pruning the BERT model on several downstream fine-tuning classification benchmarks. |
Interpretable Safety Validation for Autonomous Vehicles | An open problem for autonomous driving is how to validate the safety of an autonomous vehicle in simulation. Automated testing procedures can find failures of an autonomous system but these failures may be difficult to interpret due to their high dimensionality and may be so unlikely as to not be important. This work describes an approach for finding interpretable failures of an autonomous system. The failures are described by signal temporal logic expressions that can be understood by a human, and are optimized to produce failures that have high likelihood. Our methodology is demonstrated for the safety validation of an autonomous vehicle in the context of an unprotected left turn and a crosswalk with a pedestrian. Compared to a baseline importance sampling approach, our methodology finds more failures with higher likelihood while retaining interpretability. |