Datasets:

CShorten

/

ML-ArXiv-Papers

like 40

i-Code: An Integrative and Composable Multimodal Learning Framework

Human intelligence is multimodal; we integrate visual, linguistic, and acoustic signals to maintain a holistic worldview. Most current pretraining methods, however, are limited to one or two modalities. We present i-Code, a self-supervised pretraining framework where users may flexibly combine the modalities of vision, speech, and language into unified and general-purpose vector representations. In this framework, data from each modality are first given to pretrained single-modality encoders. The encoder outputs are then integrated with a multimodal fusion network, which uses novel attention mechanisms and other architectural innovations to effectively combine information from the different modalities. The entire system is pretrained end-to-end with new objectives including masked modality unit modeling and cross-modality contrastive learning. Unlike previous research using only video for pretraining, the i-Code framework can dynamically process single, dual, and triple-modality data during training and inference, flexibly projecting different combinations of modalities into a single representation space. Experimental results demonstrate how i-Code can outperform state-of-the-art techniques on five video understanding tasks and the GLUE NLP benchmark, improving by as much as 11% and demonstrating the power of integrative multimodal pretraining.

Diverse Image Captioning with Grounded Style

Stylized image captioning as presented in prior work aims to generate captions that reflect characteristics beyond a factual description of the scene composition, such as sentiments. Such prior work relies on given sentiment identifiers, which are used to express a certain global style in the caption, e.g. positive or negative, however without taking into account the stylistic content of the visual scene. To address this shortcoming, we first analyze the limitations of current stylized captioning datasets and propose COCO attribute-based augmentations to obtain varied stylized captions from COCO annotations. Furthermore, we encode the stylized information in the latent space of a Variational Autoencoder; specifically, we leverage extracted image attributes to explicitly structure its sequential latent space according to different localized style characteristics. Our experiments on the Senticap and COCO datasets show the ability of our approach to generate accurate captions with diversity in styles that are grounded in the image.

Assessing Dataset Bias in Computer Vision

A biased dataset is a dataset that generally has attributes with an uneven class distribution. These biases have the tendency to propagate to the models that train on them, often leading to a poor performance in the minority class. In this project, we will explore the extent to which various data augmentation methods alleviate intrinsic biases within the dataset. We will apply several augmentation techniques on a sample of the UTKFace dataset, such as undersampling, geometric transformations, variational autoencoders (VAEs), and generative adversarial networks (GANs). We then trained a classifier for each of the augmented datasets and evaluated their performance on the native test set and on external facial recognition datasets. We have also compared their performance to the state-of-the-art attribute classifier trained on the FairFace dataset. Through experimentation, we were able to find that training the model on StarGAN-generated images led to the best overall performance. We also found that training on geometrically transformed images lead to a similar performance with a much quicker training time. Additionally, the best performing models also exhibit a uniform performance across the classes within each attribute. This signifies that the model was also able to mitigate the biases present in the baseline model that was trained on the original training set. Finally, we were able to show that our model has a better overall performance and consistency on age and ethnicity classification on multiple datasets when compared with the FairFace model. Our final model has an accuracy on the UTKFace test set of 91.75%, 91.30%, and 87.20% for the gender, age, and ethnicity attribute respectively, with a standard deviation of less than 0.1 between the accuracies of the classes of each attribute.

Frequency Domain-Based Detection of Generated Audio

Attackers may manipulate audio with the intent of presenting falsified reports, changing an opinion of a public figure, and winning influence and power. The prevalence of inauthentic multimedia continues to rise, so it is imperative to develop a set of tools that determines the legitimacy of media. We present a method that analyzes audio signals to determine whether they contain real human voices or fake human voices (i.e., voices generated by neural acoustic and waveform models). Instead of analyzing the audio signals directly, the proposed approach converts the audio signals into spectrogram images displaying frequency, intensity, and temporal content and evaluates them with a Convolutional Neural Network (CNN). Trained on both genuine human voice signals and synthesized voice signals, we show our approach achieves high accuracy on this classification task.

Splicing Detection and Localization In Satellite Imagery Using Conditional GANs

The widespread availability of image editing tools and improvements in image processing techniques allow image manipulation to be very easy. Oftentimes, easy-to-use yet sophisticated image manipulation tools yields distortions/changes imperceptible to the human observer. Distribution of forged images can have drastic ramifications, especially when coupled with the speed and vastness of the Internet. Therefore, verifying image integrity poses an immense and important challenge to the digital forensic community. Satellite images specifically can be modified in a number of ways, including the insertion of objects to hide existing scenes and structures. In this paper, we describe the use of a Conditional Generative Adversarial Network (cGAN) to identify the presence of such spliced forgeries within satellite images. Additionally, we identify their locations and shapes. Trained on pristine and falsified images, our method achieves high success on these detection and localization objectives.

Exploration of the possibility of infusing Social Media Trends into generating NFT Recommendations

Recommendations Systems have been identified to be one of the integral elements of driving sales in e-commerce sites. The utilization of opinion mining data extracted from trends has been attempted to improve the recommendations that can be provided by baseline methods in this research when user-click data is lacking or is difficult to be collected due to privacy concerns. Utilizing social trends to influence the recommendations generated for a set of unique items has been explored with the use of a suggested scoring mechanism. Embracing concepts from decentralized networks that are expected to change how users interact via the internet over the next couple of decades, the suggested Recommendations System attempts to make use of multiple sources of information, applying coherent information retrieval techniques to extract probable trending items. The proposed Recommendations Architecture in the research presents a method to integrate social trends with recommendations to produce promising outputs.

Synthesized Speech Detection Using Convolutional Transformer-Based Spectrogram Analysis

Synthesized speech is common today due to the prevalence of virtual assistants, easy-to-use tools for generating and modifying speech signals, and remote work practices. Synthesized speech can also be used for nefarious purposes, including creating a purported speech signal and attributing it to someone who did not speak the content of the signal. We need methods to detect if a speech signal is synthesized. In this paper, we analyze speech signals in the form of spectrograms with a Compact Convolutional Transformer (CCT) for synthesized speech detection. A CCT utilizes a convolutional layer that introduces inductive biases and shared weights into a network, allowing a transformer architecture to perform well with fewer data samples used for training. The CCT uses an attention mechanism to incorporate information from all parts of a signal under analysis. Trained on both genuine human voice signals and synthesized human voice signals, we demonstrate that our CCT approach successfully differentiates between genuine and synthesized speech signals.

Meta-Cognition. An Inverse-Inverse Reinforcement Learning Approach for Cognitive Radars

This paper considers meta-cognitive radars in an adversarial setting. A cognitive radar optimally adapts its waveform (response) in response to maneuvers (probes) of a possibly adversarial moving target. A meta-cognitive radar is aware of the adversarial nature of the target and seeks to mitigate the adversarial target. How should the meta-cognitive radar choose its responses to sufficiently confuse the adversary trying to estimate the radar's utility function? This paper abstracts the radar's meta-cognition problem in terms of the spectra (eigenvalues) of the state and observation noise covariance matrices, and embeds the algebraic Riccati equation into an economics-based utility maximization setup. This adversarial target is an inverse reinforcement learner. By observing a noisy sequence of radar's responses (waveforms), the adversarial target uses a statistical hypothesis test to detect if the radar is a utility maximizer. In turn, the meta-cognitive radar deliberately chooses sub-optimal responses that increasing its Type-I error probability of the adversary's detector. We call this counter-adversarial step taken by the meta-cognitive radar as inverse inverse reinforcement learning (I-IRL). We illustrate the meta-cognition results of this paper via simple numerical examples. Our approach for meta-cognition in this paper is based on revealed preference theory in micro-economics and inspired by results in differential privacy and adversarial obfuscation in machine learning.

Do More Negative Samples Necessarily Hurt in Contrastive Learning?

Recent investigations in noise contrastive estimation suggest, both empirically as well as theoretically, that while having more "negative samples" in the contrastive loss improves downstream classification performance initially, beyond a threshold, it hurts downstream performance due to a "collision-coverage" trade-off. But is such a phenomenon inherent in contrastive learning? We show in a simple theoretical setting, where positive pairs are generated by sampling from the underlying latent class (introduced by Saunshi et al. (ICML 2019)), that the downstream performance of the representation optimizing the (population) contrastive loss in fact does not degrade with the number of negative samples. Along the way, we give a structural characterization of the optimal representation in our framework, for noise contrastive estimation. We also provide empirical support for our theoretical results on CIFAR-10 and CIFAR-100 datasets.

The ICML 2022 Expressive Vocalizations Workshop and Competition: Recognizing, Generating, and Personalizing Vocal Bursts

The ICML Expressive Vocalization (ExVo) Competition is focused on understanding and generating vocal bursts: laughs, gasps, cries, and other non-verbal vocalizations that are central to emotional expression and communication. ExVo 2022, includes three competition tracks using a large-scale dataset of 59,201 vocalizations from 1,702 speakers. The first, ExVo-MultiTask, requires participants to train a multi-task model to recognize expressed emotions and demographic traits from vocal bursts. The second, ExVo-Generate, requires participants to train a generative model that produces vocal bursts conveying ten different emotions. The third, ExVo-FewShot, requires participants to leverage few-shot learning incorporating speaker identity to train a model for the recognition of 10 emotions conveyed by vocal bursts. This paper describes the three tracks and provides performance measures for baseline models using state-of-the-art machine learning strategies. The baseline for each track is as follows, for ExVo-MultiTask, a combined score, computing the harmonic mean of Concordance Correlation Coefficient (CCC), Unweighted Average Recall (UAR), and inverted Mean Absolute Error (MAE) ($S_{MTL}$) is at best, 0.335 $S_{MTL}$; for ExVo-Generate, we report Fr\'echet inception distance (FID) scores ranging from 4.81 to 8.27 (depending on the emotion) between the training set and generated samples. We then combine the inverted FID with perceptual ratings of the generated samples ($S_{Gen}$) and obtain 0.174 $S_{Gen}$; and for ExVo-FewShot, a mean CCC of 0.444 is obtained.

Explain and Conquer: Personalised Text-based Reviews to Achieve Transparency

There are many contexts where dyadic data is present. Social networking is a well-known example, where transparency has grown on importance. In these contexts, pairs of items are linked building a network where interactions play a crucial role. Explaining why these relationships are established is core to address transparency. These explanations are often presented using text, thanks to the spread of the natural language understanding tasks. We have focused on the TripAdvisor platform, considering the applicability to other dyadic data contexts. The items are a subset of users and restaurants and the interactions the reviews posted by these users. Our aim is to represent and explain pairs (user, restaurant) established by agents (e.g., a recommender system or a paid promotion mechanism), so that personalisation is taken into account. We propose the PTER (Personalised TExt-based Reviews) model. We predict, from the available reviews for a given restaurant, those that fit to the specific user interactions. PTER leverages the BERT (Bidirectional Encoders Representations from Transformers) language model. We customised a deep neural network following the feature-based approach. The performance metrics show the validity of our labelling proposal. We defined an evaluation framework based on a clustering process to assess our personalised representation. PTER clearly outperforms the proposed adversary in 5 of the 6 datasets, with a minimum ratio improvement of 4%.

Differentiable Simulation of Soft Multi-body Systems

We present a method for differentiable simulation of soft articulated bodies. Our work enables the integration of differentiable physical dynamics into gradient-based pipelines. We develop a top-down matrix assembly algorithm within Projective Dynamics and derive a generalized dry friction model for soft continuum using a new matrix splitting strategy. We derive a differentiable control framework for soft articulated bodies driven by muscles, joint torques, or pneumatic tubes. The experiments demonstrate that our designs make soft body simulation more stable and realistic compared to other frameworks. Our method accelerates the solution of system identification problems by more than an order of magnitude, and enables efficient gradient-based learning of motion control with soft robots.

XLTime: A Cross-Lingual Knowledge Transfer Framework for Temporal Expression Extraction

Temporal Expression Extraction (TEE) is essential for understanding time in natural language. It has applications in Natural Language Processing (NLP) tasks such as question answering, information retrieval, and causal inference. To date, work in this area has mostly focused on English as there is a scarcity of labeled data for other languages. We propose XLTime, a novel framework for multilingual TEE. XLTime works on top of pre-trained language models and leverages multi-task learning to prompt cross-language knowledge transfer both from English and within the non-English languages. XLTime alleviates problems caused by a shortage of data in the target language. We apply XLTime with different language models and show that it outperforms the previous automatic SOTA methods on French, Spanish, Portuguese, and Basque, by large margins. XLTime also closes the gap considerably on the handcrafted HeidelTime method.

Bézier Curve Gaussian Processes

Probabilistic models for sequential data are the basis for a variety of applications concerned with processing timely ordered information. The predominant approach in this domain is given by neural networks, which incorporate either stochastic units or components. This paper proposes a new probabilistic sequence model building on probabilistic B\'ezier curves. Using Gaussian distributed control points, these parametric curves pose a special case for Gaussian processes (GP). Combined with a Mixture Density network, Bayesian conditional inference can be performed without the need for mean field variational approximation or Monte Carlo simulation, which is a requirement of common approaches. For assessing this hybrid model's viability, it is applied to an exemplary sequence prediction task. In this case the model is used for pedestrian trajectory prediction, where a generated prediction also serves as a GP prior. Following this, the initial prediction can be refined using the GP framework by calculating different posterior distributions, in order to adapt more towards a given observed trajectory segment.

Self-focusing virtual screening with active design space pruning

High-throughput virtual screening is an indispensable technique utilized in the discovery of small molecules. In cases where the library of molecules is exceedingly large, the cost of an exhaustive virtual screen may be prohibitive. Model-guided optimization has been employed to lower these costs through dramatic increases in sample efficiency compared to random selection. However, these techniques introduce new costs to the workflow through the surrogate model training and inference steps. In this study, we propose an extension to the framework of model-guided optimization that mitigates inferences costs using a technique we refer to as design space pruning (DSP), which irreversibly removes poor-performing candidates from consideration. We study the application of DSP to a variety of optimization tasks and observe significant reductions in overhead costs while exhibiting similar performance to the baseline optimization. DSP represents an attractive extension of model-guided optimization that can limit overhead costs in optimization settings where these costs are non-negligible relative to objective costs, such as docking.

Detection of Propaganda Techniques in Visuo-Lingual Metaphor in Memes

The exponential rise of social media networks has allowed the production, distribution, and consumption of data at a phenomenal rate. Moreover, the social media revolution has brought a unique phenomenon to social media platforms called Internet memes. Internet memes are one of the most popular contents used on social media, and they can be in the form of images with a witty, catchy, or satirical text description. In this paper, we are dealing with propaganda that is often seen in Internet memes in recent times. Propaganda is communication, which frequently includes psychological and rhetorical techniques to manipulate or influence an audience to act or respond as the propagandist wants. To detect propaganda in Internet memes, we propose a multimodal deep learning fusion system that fuses the text and image feature representations and outperforms individual models based solely on either text or image modalities.

Don't sweat the small stuff, classify the rest: Sample Shielding to protect text classifiers against adversarial attacks

Deep learning (DL) is being used extensively for text classification. However, researchers have demonstrated the vulnerability of such classifiers to adversarial attacks. Attackers modify the text in a way which misleads the classifier while keeping the original meaning close to intact. State-of-the-art (SOTA) attack algorithms follow the general principle of making minimal changes to the text so as to not jeopardize semantics. Taking advantage of this we propose a novel and intuitive defense strategy called Sample Shielding. It is attacker and classifier agnostic, does not require any reconfiguration of the classifier or external resources and is simple to implement. Essentially, we sample subsets of the input text, classify them and summarize these into a final decision. We shield three popular DL text classifiers with Sample Shielding, test their resilience against four SOTA attackers across three datasets in a realistic threat setting. Even when given the advantage of knowing about our shielding strategy the adversary's attack success rate is <=10% with only one exception and often < 5%. Additionally, Sample Shielding maintains near original accuracy when applied to original texts. Crucially, we show that the `make minimal changes' approach of SOTA attackers leads to critical vulnerabilities that can be defended against with an intuitive sampling strategy.

MemSE: Fast MSE Prediction for Noisy Memristor-Based DNN Accelerators

Memristors enable the computation of matrix-vector multiplications (MVM) in memory and, therefore, show great potential in highly increasing the energy efficiency of deep neural network (DNN) inference accelerators. However, computations in memristors suffer from hardware non-idealities and are subject to different sources of noise that may negatively impact system performance. In this work, we theoretically analyze the mean squared error of DNNs that use memristor crossbars to compute MVM. We take into account both the quantization noise, due to the necessity of reducing the DNN model size, and the programming noise, stemming from the variability during the programming of the memristance value. Simulations on pre-trained DNN models showcase the accuracy of the analytical prediction. Furthermore the proposed method is almost two order of magnitude faster than Monte-Carlo simulation, thus making it possible to optimize the implementation parameters to achieve minimal error for a given power constraint.

On Circuit Depth Scaling For Quantum Approximate Optimization

Variational quantum algorithms are the centerpiece of modern quantum programming. These algorithms involve training parameterized quantum circuits using a classical co-processor, an approach adapted partly from classical machine learning. An important subclass of these algorithms, designed for combinatorial optimization on currrent quantum hardware, is the quantum approximate optimization algorithm (QAOA). It is known that problem density - a problem constraint to variable ratio - induces under-parametrization in fixed depth QAOA. Density dependent performance has been reported in the literature, yet the circuit depth required to achieve fixed performance (henceforth called critical depth) remained unknown. Here, we propose a predictive model, based on a logistic saturation conjecture for critical depth scaling with respect to density. Focusing on random instances of MAX-2-SAT, we test our predictive model against simulated data with up to 15 qubits. We report the average critical depth, required to attain a success probability of 0.7, saturates at a value of 10 for densities beyond 4. We observe the predictive model to describe the simulated data within a $3\sigma$ confidence interval. Furthermore, based on the model, a linear trend for the critical depth with respect problem size is recovered for the range of 5 to 15 qubits.

Adversarial Training for High-Stakes Reliability

In the future, powerful AI systems may be deployed in high-stakes settings, where a single failure could be catastrophic. One technique for improving AI safety in high-stakes settings is adversarial training, which uses an adversary to generate examples to train on in order to achieve better worst-case performance. In this work, we used a language generation task as a testbed for achieving high reliability through adversarial training. We created a series of adversarial training techniques -- including a tool that assists human adversaries -- to find and eliminate failures in a classifier that filters text completions suggested by a generator. In our simple "avoid injuries" task, we determined that we can set very conservative classifier thresholds without significantly impacting the quality of the filtered outputs. With our chosen thresholds, filtering with our baseline classifier decreases the rate of unsafe completions from about 2.4% to 0.003% on in-distribution data, which is near the limit of our ability to measure. We found that adversarial training significantly increased robustness to the adversarial attacks that we trained on, without affecting in-distribution performance. We hope to see further work in the high-stakes reliability setting, including more powerful tools for enhancing human adversaries and better ways to measure high levels of reliability, until we can confidently rule out the possibility of catastrophic deployment-time failures of powerful models.

Deep Sequence Modeling for Anomalous ISP Traffic Prediction

Internet traffic in the real world is susceptible to various external and internal factors which may abruptly change the normal traffic flow. Those unexpected changes are considered outliers in traffic. However, deep sequence models have been used to predict complex IP traffic, but their comparative performance for anomalous traffic has not been studied extensively. In this paper, we investigated and evaluated the performance of different deep sequence models for anomalous traffic prediction. Several deep sequences models were implemented to predict real traffic without and with outliers and show the significance of outlier detection in real-world traffic prediction. First, two different outlier detection techniques, such as the Three-Sigma rule and Isolation Forest, were applied to identify the anomaly. Second, we adjusted those abnormal data points using the Backward Filling technique before training the model. Finally, the performance of different models was compared for abnormal and adjusted traffic. LSTM_Encoder_Decoder (LSTM_En_De) is the best prediction model in our experiment, reducing the deviation between actual and predicted traffic by more than 11\% after adjusting the outliers. All other models, including Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), LSTM_En_De with Attention layer (LSTM_En_De_Atn), Gated Recurrent Unit (GRU), show better prediction after replacing the outliers and decreasing prediction error by more than 29%, 24%, 19%, and 10% respectively. Our experimental results indicate that the outliers in the data can significantly impact the quality of the prediction. Thus, outlier detection and mitigation assist the deep sequence model in learning the general trend and making better predictions.

Automatic Segmentation of Aircraft Dents in Point Clouds

Dents on the aircraft skin are frequent and may easily go undetected during airworthiness checks, as their inspection process is tedious and extremely subject to human factors and environmental conditions. Nowadays, 3D scanning technologies are being proposed for more reliable, human-independent measurements, yet the process of inspection and reporting remains laborious and time consuming because data acquisition and validation are still carried out by the engineer. For full automation of dent inspection, the acquired point cloud data must be analysed via a reliable segmentation algorithm, releasing humans from the search and evaluation of damage. This paper reports on two developments towards automated dent inspection. The first is a method to generate a synthetic dataset of dented surfaces to train a fully convolutional neural network. The training of machine learning algorithms needs a substantial volume of dent data, which is not readily available. Dents are thus simulated in random positions and shapes, within criteria and definitions of a Boeing 737 structural repair manual. The noise distribution from the scanning apparatus is then added to reflect the complete process of 3D point acquisition on the training. The second proposition is a surface fitting strategy to convert 3D point clouds to 2.5D. This allows higher resolution point clouds to be processed with a small amount of memory compared with state-of-the-art methods involving 3D sampling approaches. Simulations with available ground truth data show that the proposed technique reaches an intersection-over-union of over 80%. Experiments over dent samples prove an effective detection of dents with a speed of over 500 000 points per second.

Local Stochastic Bilevel Optimization with Momentum-Based Variance Reduction

Bilevel Optimization has witnessed notable progress recently with new emerging efficient algorithms and has been applied to many machine learning tasks such as data cleaning, few-shot learning, and neural architecture search. However, little attention has been paid to solve the bilevel problems under distributed setting. Federated learning (FL) is an emerging paradigm which solves machine learning tasks over distributed-located data. FL problems are challenging to solve due to the heterogeneity and communication bottleneck. However, it is unclear how these challenges will affect the convergence of Bilevel Optimization algorithms. In this paper, we study Federated Bilevel Optimization problems. Specifically, we first propose the FedBiO, a deterministic gradient-based algorithm and we show it requires $O(\epsilon^{-2})$ number of iterations to reach an $\epsilon$-stationary point. Then we propose FedBiOAcc to accelerate FedBiO with the momentum-based variance-reduction technique under the stochastic scenario. We show FedBiOAcc has complexity of $O(\epsilon^{-1.5})$. Finally, we validate our proposed algorithms via the important Fair Federated Learning task. More specifically, we define a bilevel-based group fair FL objective. Our algorithms show superior performances compared to other baselines in numerical experiments.

Modeling and Correcting Bias in Sequential Evaluation

We consider the problem of sequential evaluation, in which an evaluator observes candidates in a sequence and assigns scores to these candidates in an online, irrevocable fashion. Motivated by the psychology literature that has studied sequential bias in such settings -- namely, dependencies between the evaluation outcome and the order in which the candidates appear -- we propose a natural model for the evaluator's rating process that captures the lack of calibration inherent to such a task. We conduct crowdsourcing experiments to demonstrate various facets of our model. We then proceed to study how to correct sequential bias under our model by posing this as a statistical inference problem. We propose a near-linear time, online algorithm for this task and prove guarantees in terms of two canonical ranking metrics, matched with lower bounds demonstrating optimality in a certain sense. Our algorithm outperforms the de facto method of using the rankings induced by the reported scores.

An Empirical Study on Internet Traffic Prediction Using Statistical Rolling Model

Real-world IP network traffic is susceptible to external and internal factors such as new internet service integration, traffic migration, internet application, etc. Due to these factors, the actual internet traffic is non-linear and challenging to analyze using a statistical model for future prediction. In this paper, we investigated and evaluated the performance of different statistical prediction models for real IP network traffic; and showed a significant improvement in prediction using the rolling prediction technique. Initially, a set of best hyper-parameters for the corresponding prediction model is identified by analyzing the traffic characteristics and implementing a grid search algorithm based on the minimum Akaike Information Criterion (AIC). Then, we performed a comparative performance analysis among AutoRegressive Integrated Moving Average (ARIMA), Seasonal ARIMA (SARIMA), SARIMA with eXogenous factors (SARIMAX), and Holt-Winter for single-step prediction. The seasonality of our traffic has been explicitly modeled using SARIMA, which reduces the rolling prediction Mean Average Percentage Error (MAPE) by more than 4% compared to ARIMA (incapable of handling the seasonality). We further improved traffic prediction using SARIMAX to learn different exogenous factors extracted from the original traffic, which yielded the best rolling prediction results with a MAPE of 6.83%. Finally, we applied the exponential smoothing technique to handle the variability in traffic following the Holt-Winter model, which exhibited a better prediction than ARIMA (around 1.5% less MAPE). The rolling prediction technique reduced prediction error using real Internet Service Provider (ISP) traffic data by more than 50\% compared to the standard prediction method.

RAFT-MSF: Self-Supervised Monocular Scene Flow using Recurrent Optimizer

Learning scene flow from a monocular camera still remains a challenging task due to its ill-posedness as well as lack of annotated data. Self-supervised methods demonstrate learning scene flow estimation from unlabeled data, yet their accuracy lags behind (semi-)supervised methods. In this paper, we introduce a self-supervised monocular scene flow method that substantially improves the accuracy over the previous approaches. Based on RAFT, a state-of-the-art optical flow model, we design a new decoder to iteratively update 3D motion fields and disparity maps simultaneously. Furthermore, we propose an enhanced upsampling layer and a disparity initialization technique, which overall further improves accuracy up to 7.2%. Our method achieves state-of-the-art accuracy among all self-supervised monocular scene flow methods, improving accuracy by 34.2%. Our fine-tuned model outperforms the best previous semi-supervised method with 228 times faster runtime. Code will be publicly available.

Privacy Amplification via Random Participation in Federated Learning

Running a randomized algorithm on a subsampled dataset instead of the entire dataset amplifies differential privacy guarantees. In this work, in a federated setting, we consider random participation of the clients in addition to subsampling their local datasets. Since such random participation of the clients creates correlation among the samples of the same client in their subsampling, we analyze the corresponding privacy amplification via non-uniform subsampling. We show that when the size of the local datasets is small, the privacy guarantees via random participation is close to those of the centralized setting, in which the entire dataset is located in a single host and subsampled. On the other hand, when the local datasets are large, observing the output of the algorithm may disclose the identities of the sampled clients with high confidence. Our analysis reveals that, even in this case, privacy guarantees via random participation outperform those via only local subsampling.

Efficient Fine-Tuning of BERT Models on the Edge

Resource-constrained devices are increasingly the deployment targets of machine learning applications. Static models, however, do not always suffice for dynamic environments. On-device training of models allows for quick adaptability to new scenarios. With the increasing size of deep neural networks, as noted with the likes of BERT and other natural language processing models, comes increased resource requirements, namely memory, computation, energy, and time. Furthermore, training is far more resource intensive than inference. Resource-constrained on-device learning is thus doubly difficult, especially with large BERT-like models. By reducing the memory usage of fine-tuning, pre-trained BERT models can become efficient enough to fine-tune on resource-constrained devices. We propose Freeze And Reconfigure (FAR), a memory-efficient training regime for BERT-like models that reduces the memory usage of activation maps during fine-tuning by avoiding unnecessary parameter updates. FAR reduces fine-tuning time on the DistilBERT model and CoLA dataset by 30%, and time spent on memory operations by 47%. More broadly, reductions in metric performance on the GLUE and SQuAD datasets are around 1% on average.

BiOcularGAN: Bimodal Synthesis and Annotation of Ocular Images

Current state-of-the-art segmentation techniques for ocular images are critically dependent on large-scale annotated datasets, which are labor-intensive to gather and often raise privacy concerns. In this paper, we present a novel framework, called BiOcularGAN, capable of generating synthetic large-scale datasets of photorealistic (visible light and near infrared) ocular images, together with corresponding segmentation labels to address these issues. At its core, the framework relies on a novel Dual-Branch StyleGAN2 (DB-StyleGAN2) model that facilitates bimodal image generation, and a Semantic Mask Generator (SMG) that produces semantic annotations by exploiting DB-StyleGAN2's feature space. We evaluate BiOcularGAN through extensive experiments across five diverse ocular datasets and analyze the effects of bimodal data generation on image quality and the produced annotations. Our experimental results show that BiOcularGAN is able to produce high-quality matching bimodal images and annotations (with minimal manual intervention) that can be used to train highly competitive (deep) segmentation models that perform well across multiple real-world datasets. The source code will be made publicly available.

Explainable multi-class anomaly detection on functional data

In this paper we describe an approach for anomaly detection and its explainability in multivariate functional data. The anomaly detection procedure consists of transforming the series into a vector of features and using an Isolation forest algorithm. The explainable procedure is based on the computation of the SHAP coefficients and on the use of a supervised decision tree. We apply it on simulated data to measure the performance of our method and on real data coming from industry.

ExSpliNet: An interpretable and expressive spline-based neural network

In this paper we present ExSpliNet, an interpretable and expressive neural network model. The model combines ideas of Kolmogorov neural networks, ensembles of probabilistic trees, and multivariate B-spline representations. We give a probabilistic interpretation of the model and show its universal approximation properties. We also discuss how it can be efficiently encoded by exploiting B-spline properties. Finally, we test the effectiveness of the proposed model on synthetic approximation problems and classical machine learning benchmark datasets.

Compact Neural Networks via Stacking Designed Basic Units

Unstructured pruning has the limitation of dealing with the sparse and irregular weights. By contrast, structured pruning can help eliminate this drawback but it requires complex criterion to determine which components to be pruned. To this end, this paper presents a new method termed TissueNet, which directly constructs compact neural networks with fewer weight parameters by independently stacking designed basic units, without requiring additional judgement criteria anymore. Given the basic units of various architectures, they are combined and stacked in a certain form to build up compact neural networks. We formulate TissueNet in diverse popular backbones for comparison with the state-of-the-art pruning methods on different benchmark datasets. Moreover, two new metrics are proposed to evaluate compression performance. Experiment results show that TissueNet can achieve comparable classification accuracy while saving up to around 80% FLOPs and 89.7% parameters. That is, stacking basic units provides a new promising way for network compression.

Meta Learning for Natural Language Processing: A Survey

Deep learning has been the mainstream technique in natural language processing (NLP) area. However, the techniques require many labeled data and are less generalizable across domains. Meta-learning is an arising field in machine learning studying approaches to learn better learning algorithms. Approaches aim at improving algorithms in various aspects, including data efficiency and generalizability. Efficacy of approaches has been shown in many NLP tasks, but there is no systematic survey of these approaches in NLP, which hinders more researchers from joining the field. Our goal with this survey paper is to offer researchers pointers to relevant meta-learning works in NLP and attract more attention from the NLP community to drive future innovation. This paper first introduces the general concepts of meta-learning and the common approaches. Then we summarize task construction settings and application of meta-learning for various NLP problems and review the development of meta-learning in NLP community.

On the uncertainty principle of neural networks

Despite the successes in many fields, it is found that neural networks are vulnerability and difficult to be both accurate and robust (robust means that the prediction of the trained network stays unchanged for inputs with non-random perturbations introduced by adversarial attacks). Various empirical and analytic studies have suggested that there is more or less a trade-off between the accuracy and robustness of neural networks. If the trade-off is inherent, applications based on the neural networks are vulnerable with untrustworthy predictions. It is then essential to ask whether the trade-off is an inherent property or not. Here, we show that the accuracy-robustness trade-off is an intrinsic property whose underlying mechanism is deeply related to the uncertainty principle in quantum mechanics. We find that for a neural network to be both accurate and robust, it needs to resolve the features of the two conjugated parts $x$ (the inputs) and $\Delta$ (the derivatives of the normalized loss function $J$ with respect to $x$), respectively. Analogous to the position-momentum conjugation in quantum mechanics, we show that the inputs and their conjugates cannot be resolved by a neural network simultaneously.

A unified view on Self-Organizing Maps (SOMs) and Stochastic Neighbor Embedding (SNE)

We propose a unified view on two widely used data visualization techniques: Self-Organizing Maps (SOMs) and Stochastic Neighbor Embedding (SNE). We show that they can both be derived from a common mathematical framework. Leveraging this formulation, we propose to compare SOM and SNE quantitatively on two datasets, and discuss possible avenues for future work to take advantage of both approaches.

Subspace Diffusion Generative Models

Score-based models generate samples by mapping noise to data (and vice versa) via a high-dimensional diffusion process. We question whether it is necessary to run this entire process at high dimensionality and incur all the inconveniences thereof. Instead, we restrict the diffusion via projections onto subspaces as the data distribution evolves toward noise. When applied to state-of-the-art models, our framework simultaneously improves sample quality -- reaching an FID of 2.17 on unconditional CIFAR-10 -- and reduces the computational cost of inference for the same number of denoising steps. Our framework is fully compatible with continuous-time diffusion and retains its flexible capabilities, including exact log-likelihoods and controllable generation. Code is available at https://github.com/bjing2016/subspace-diffusion.

Scalable Regularised Joint Mixture Models

In many applications, data can be heterogeneous in the sense of spanning latent groups with different underlying distributions. When predictive models are applied to such data the heterogeneity can affect both predictive performance and interpretability. Building on developments at the intersection of unsupervised learning and regularised regression, we propose an approach for heterogeneous data that allows joint learning of (i) explicit multivariate feature distributions, (ii) high-dimensional regression models and (iii) latent group labels, with both (i) and (ii) specific to latent groups and both elements informing (iii). The approach is demonstrably effective in high dimensions, combining data reduction for computational efficiency with a re-weighting scheme that retains key signals even when the number of features is large. We discuss in detail these aspects and their impact on modelling and computation, including EM convergence. The approach is modular and allows incorporation of data reductions and high-dimensional estimators that are suitable for specific applications. We show results from extensive simulations and real data experiments, including highly non-Gaussian data. Our results allow efficient, effective analysis of high-dimensional data in settings, such as biomedicine, where both interpretable prediction and explicit feature space models are needed but hidden heterogeneity may be a concern.

Residual Graph Convolutional Recurrent Networks For Multi-step Traffic Flow Forecasting

Traffic flow forecasting is essential for traffic planning, control and management. The main challenge of traffic forecasting tasks is accurately capturing traffic networks' spatial and temporal correlation. Although there are many traffic forecasting methods, most of them still have limitations in capturing spatial and temporal correlations. To improve traffic forecasting accuracy, we propose a new Spatial-temporal forecasting model, namely the Residual Graph Convolutional Recurrent Network (RGCRN). The model uses our proposed Residual Graph Convolutional Network (ResGCN) to capture the fine-grained spatial correlation of the traffic road network and then uses a Bi-directional Gated Recurrent Unit (BiGRU) to model time series with spatial information and obtains the temporal correlation by analysing the change in information transfer between the forward and reverse neurons of the time series data. Our comparative experimental results on two real datasets show that RGCRN improves on average by 20.66% compared to the best baseline model. You can get our source code and data through https://github.com/zhangshqii/RGCRN.

Revisiting Communication-Efficient Federated Learning with Balanced Global and Local Updates

In federated learning (FL), a number of devices train their local models and upload the corresponding parameters or gradients to the base station (BS) to update the global model while protecting their data privacy. However, due to the limited computation and communication resources, the number of local trainings (a.k.a. local update) and that of aggregations (a.k.a. global update) need to be carefully chosen. In this paper, we investigate and analyze the optimal trade-off between the number of local trainings and that of global aggregations to speed up the convergence and enhance the prediction accuracy over the existing works. Our goal is to minimize the global loss function under both the delay and the energy consumption constraints. In order to make the optimization problem tractable, we derive a new and tight upper bound on the loss function, which allows us to obtain closed-form expressions for the number of local trainings and that of global aggregations. Simulation results show that our proposed scheme can achieve a better performance in terms of the prediction accuracy, and converge much faster than the baseline schemes.

On the Convergence of Fictitious Play: A Decomposition Approach

Fictitious play (FP) is one of the most fundamental game-theoretical learning frameworks for computing Nash equilibrium in $n$-player games, which builds the foundation for modern multi-agent learning algorithms. Although FP has provable convergence guarantees on zero-sum games and potential games, many real-world problems are often a mixture of both and the convergence property of FP has not been fully studied yet. In this paper, we extend the convergence results of FP to the combinations of such games and beyond. Specifically, we derive new conditions for FP to converge by leveraging game decomposition techniques. We further develop a linear relationship unifying cooperation and competition in the sense that these two classes of games are mutually transferable. Finally, we analyze a non-convergent example of FP, the Shapley game, and develop sufficient conditions for FP to converge.

Efficient implementation of incremental proximal-point methods

Model training algorithms which observe a small portion of the training set in each computational step are ubiquitous in practical machine learning, and include both stochastic and online optimization methods. In the vast majority of cases, such algorithms typically observe the training samples via the gradients of the cost functions the samples incur. Thus, these methods exploit are the \emph{slope} of the cost functions via their first-order approximations. To address limitations of gradient-based methods, such as sensitivity to step-size choice in the stochastic setting, or inability to exploit small function variability in the online setting, several streams of research attempt to exploit more information about the cost functions than just their gradients via the well-known proximal framework of optimization. However, implementing such methods in practice poses a challenge, since each iteration step boils down to computing a proximal operator, which may not be easy. In this work we provide efficient algorithms and corresponding implementations of proximal operators in order to make experimentation with incremental proximal optimization algorithms accessible to a larger audience of researchers and practitioners, and in particular to promote additional theoretical research into these methods by closing the gap between their theoretical description in research papers and their use in practice. The corresponding code is published at https://github.com/alexshtf/inc_prox_pt.

Predicting vacant parking space availability zone-wisely: a graph based spatio-temporal prediction approach

Vacant parking space (VPS) prediction is one of the key issues of intelligent parking guidance systems. Accurately predicting VPS information plays a crucial role in intelligent parking guidance systems, which can help drivers find parking space quickly, reducing unnecessary waste of time and excessive environmental pollution. Through the simple analysis of historical data, we found that there not only exists a obvious temporal correlation in each parking lot, but also a clear spatial correlation between different parking lots. In view of this, this paper proposed a graph data-based model ST-GBGRU (Spatial-Temporal Graph Based Gated Recurrent Unit), the number of VPSs can be predicted both in short-term (i.e., within 30 min) and in long-term (i.e., over 30min). On the one hand, the temporal correlation of historical VPS data is extracted by GRU, on the other hand, the spatial correlation of historical VPS data is extracted by GCN inside GRU. Two prediction methods, namely direct prediction and iterative prediction, are combined with the proposed model. Finally, the prediction model is applied to predict the number VPSs of 8 public parking lots in Santa Monica. The results show that in the short-term and long-term prediction tasks, ST-GBGRU model can achieve high accuracy and have good application prospects.

High-dimensional Asymptotics of Feature Learning: How One Gradient Step Improves the Representation

We study the first gradient descent step on the first-layer parameters $\boldsymbol{W}$ in a two-layer neural network: $f(\boldsymbol{x}) = \frac{1}{\sqrt{N}}\boldsymbol{a}^\top\sigma(\boldsymbol{W}^\top\boldsymbol{x})$, where $\boldsymbol{W}\in\mathbb{R}^{d\times N}, \boldsymbol{a}\in\mathbb{R}^{N}$ are randomly initialized, and the training objective is the empirical MSE loss: $\frac{1}{n}\sum_{i=1}^n (f(\boldsymbol{x}_i)-y_i)^2$. In the proportional asymptotic limit where $n,d,N\to\infty$ at the same rate, and an idealized student-teacher setting, we show that the first gradient update contains a rank-1 "spike", which results in an alignment between the first-layer weights and the linear component of the teacher model $f^*$. To characterize the impact of this alignment, we compute the prediction risk of ridge regression on the conjugate kernel after one gradient step on $\boldsymbol{W}$ with learning rate $\eta$, when $f^*$ is a single-index model. We consider two scalings of the first step learning rate $\eta$. For small $\eta$, we establish a Gaussian equivalence property for the trained feature map, and prove that the learned kernel improves upon the initial random features model, but cannot defeat the best linear model on the input. Whereas for sufficiently large $\eta$, we prove that for certain $f^*$, the same ridge estimator on trained features can go beyond this "linear regime" and outperform a wide range of random features and rotationally invariant kernels. Our results demonstrate that even one gradient step can lead to a considerable advantage over random features, and highlight the role of learning rate scaling in the initial phase of training.

Learning Coulomb Diamonds in Large Quantum Dot Arrays

We introduce an algorithm that is able to find the facets of Coulomb diamonds in quantum dot arrays. We simulate these arrays using the constant-interaction model, and rely only on one-dimensional raster scans (rays) to learn a model of the device using regularized maximum likelihood estimation. This allows us to determine, for a given charge state of the device, which transitions exist and what the compensated gate voltages for these are. For smaller devices the simulator can also be used to compute the exact boundaries of the Coulomb diamonds, which we use to assess that our algorithm correctly finds the vast majority of transitions with high precision.

Efficient and Convergent Federated Learning

Federated learning has shown its advances over the last few years but is facing many challenges, such as how algorithms save communication resources, how they reduce computational costs, and whether they converge. To address these issues, this paper proposes a new federated learning algorithm (FedGiA) that combines the gradient descent and the inexact alternating direction method of multipliers. It is shown that FedGiA is computation and communication-efficient and convergent linearly under mild conditions.

RLFlow: Optimising Neural Network Subgraph Transformation with World Models

Training deep learning models takes an extremely long execution time and consumes large amounts of computing resources. At the same time, recent research proposed systems and compilers that are expected to decrease deep learning models runtime. An effective optimisation methodology in data processing is desirable, and the reduction of compute requirements of deep learning models is the focus of extensive research. In this paper, we address the neural network sub-graph transformation by exploring reinforcement learning (RL) agents to achieve performance improvement. Our proposed approach RLFlow can learn to perform neural network subgraph transformations, without the need for expertly designed heuristics to achieve a high level of performance. Recent work has aimed at applying RL to computer systems with some success, especially using model-free RL techniques. Model-based reinforcement learning methods have seen an increased focus in research as they can be used to learn the transition dynamics of the environment; this can be leveraged to train an agent using a hallucinogenic environment such as World Model (WM), thereby increasing sample efficiency compared to model-free approaches. WM uses variational auto-encoders and it builds a model of the system and allows exploring the model in an inexpensive way. In RLFlow, we propose a design for a model-based agent with WM which learns to optimise the architecture of neural networks by performing a sequence of sub-graph transformations to reduce model runtime. We show that our approach can match the state-of-the-art performance on common convolutional networks and outperforms by up to 5% those based on transformer-style architectures

ARCADE: Adversarially Regularized Convolutional Autoencoder for Network Anomaly Detection

As the number of heterogenous IP-connected devices and traffic volume increase, so does the potential for security breaches. The undetected exploitation of these breaches can bring severe cybersecurity and privacy risks. In this paper, we present a practical unsupervised anomaly-based deep learning detection system called ARCADE (Adversarially Regularized Convolutional Autoencoder for unsupervised network anomaly DEtection). ARCADE exploits the property of 1D Convolutional Neural Networks (CNNs) and Generative Adversarial Networks (GAN) to automatically build a profile of the normal traffic based on a subset of raw bytes of a few initial packets of network flows so that potential network anomalies and intrusions can be effectively detected before they could cause any more damage to the network. A convolutional Autoencoder (AE) is proposed that suits online detection in resource-constrained environments, and can be easily improved for environments with higher computational capabilities. An adversarial training strategy is proposed to regularize and decrease the AE's capabilities to reconstruct network flows that are out of the normal distribution, and thereby improve its anomaly detection capabilities. The proposed approach is more effective than existing state-of-the-art deep learning approaches for network anomaly detection and significantly reduces detection time. The evaluation results show that the proposed approach is suitable for anomaly detection on resource-constrained hardware platforms such as Raspberry Pi.

Time Shifts to Reduce the Size of Reservoir Computers

A reservoir computer is a type of dynamical system arranged to do computation. Typically, a reservoir computer is constructed by connecting a large number of nonlinear nodes in a network that includes recurrent connections. In order to achieve accurate results, the reservoir usually contains hundreds to thousands of nodes. This high dimensionality makes it difficult to analyze the reservoir computer using tools from dynamical systems theory. Additionally, the need to create and connect large numbers of nonlinear nodes makes it difficult to design and build analog reservoir computers that can be faster and consume less power than digital reservoir computers. We demonstrate here that a reservoir computer may be divided into two parts; a small set of nonlinear nodes (the reservoir), and a separate set of time-shifted reservoir output signals. The time-shifted output signals serve to increase the rank and memory of the reservoir computer, and the set of nonlinear nodes may create an embedding of the input dynamical system. We use this time-shifting technique to obtain excellent performance from an opto-electronic delay-based reservoir computer with only a small number of virtual nodes. Because only a few nonlinear nodes are required, construction of a reservoir computer becomes much easier, and delay-based reservoir computers can operate at much higher speeds.

Growing Isotropic Neural Cellular Automata

Modeling the ability of multicellular organisms to build and maintain their bodies through local interactions between individual cells (morphogenesis) is a long-standing challenge of developmental biology. Recently, the Neural Cellular Automata (NCA) model was proposed as a way to find local system rules that produce a desired global behaviour, such as growing and persisting a predefined target pattern, by repeatedly applying the same rule over a grid starting from a single cell. In this work, we argue that the original Growing NCA model has an important limitation: anisotropy of the learned update rule. This implies the presence of an external factor that orients the cells in a particular direction. In other words, "physical" rules of the underlying system are not invariant to rotation, thus prohibiting the existence of differently oriented instances of the target pattern on the same grid. We propose a modified Isotropic NCA (IsoNCA) model that does not have this limitation. We demonstrate that such cell systems can be trained to grow accurate asymmetrical patterns through either of two methods: (1) by breaking symmetries using structured seeds or (2) by introducing a rotation-reflection invariant training objective and relying on symmetry-breaking caused by asynchronous cell updates.

Autonomy and Intelligence in the Computing Continuum: Challenges, Enablers, and Future Directions for Orchestration

Future AI applications require performance, reliability and privacy that the existing, cloud-dependant system architectures cannot provide. In this article, we study orchestration in the device-edge-cloud continuum, and focus on AI for edge, that is, the AI methods used in resource orchestration. We claim that to support the constantly growing requirements of intelligent applications in the device-edge-cloud computing continuum, resource orchestration needs to embrace edge AI and emphasize local autonomy and intelligence. To justify the claim, we provide a general definition for continuum orchestration, and look at how current and emerging orchestration paradigms are suitable for the computing continuum. We describe certain major emerging research themes that may affect future orchestration, and provide an early vision of an orchestration paradigm that embraces those research themes. Finally, we survey current key edge AI methods and look at how they may contribute into fulfilling the vision of future continuum orchestration.

A Falsificationist Account of Artificial Neural Networks

Machine learning operates at the intersection of statistics and computer science. This raises the question as to its underlying methodology. While much emphasis has been put on the close link between the process of learning from data and induction, the falsificationist component of machine learning has received minor attention. In this paper, we argue that the idea of falsification is central to the methodology of machine learning. It is commonly thought that machine learning algorithms infer general prediction rules from past observations. This is akin to a statistical procedure by which estimates are obtained from a sample of data. But machine learning algorithms can also be described as choosing one prediction rule from an entire class of functions. In particular, the algorithm that determines the weights of an artificial neural network operates by empirical risk minimization and rejects prediction rules that lack empirical adequacy. It also exhibits a behavior of implicit regularization that pushes hypothesis choice toward simpler prediction rules. We argue that taking both aspects together gives rise to a falsificationist account of artificial neural networks.

An Empirical Analysis of the Use of Real-Time Reachability for the Safety Assurance of Autonomous Vehicles

Recent advances in machine learning technologies and sensing have paved the way for the belief that safe, accessible, and convenient autonomous vehicles may be realized in the near future. Despite tremendous advances within this context, fundamental challenges around safety and reliability are limiting their arrival and comprehensive adoption. Autonomous vehicles are often tasked with operating in dynamic and uncertain environments. As a result, they often make use of highly complex components, such as machine learning approaches, to handle the nuances of sensing, actuation, and control. While these methods are highly effective, they are notoriously difficult to assure. Moreover, within uncertain and dynamic environments, design time assurance analyses may not be sufficient to guarantee safety. Thus, it is critical to monitor the correctness of these systems at runtime. One approach for providing runtime assurance of systems with components that may not be amenable to formal analysis is the simplex architecture, where an unverified component is wrapped with a safety controller and a switching logic designed to prevent dangerous behavior. In this paper, we propose using a real-time reachability algorithm for the implementation of the simplex architecture to assure the safety of a 1/10 scale open source autonomous vehicle platform known as F1/10. The reachability algorithm that we leverage (a) provides provable guarantees of safety, and (b) is used to detect potentially unsafe scenarios. In our approach, the need to analyze an underlying controller is abstracted away, instead focusing on the effects of the controller's decisions on the system's future states. We demonstrate the efficacy of our architecture through a vast set of experiments conducted both in simulation and on an embedded hardware platform.

Multimodal Detection of Unknown Objects on Roads for Autonomous Driving

Tremendous progress in deep learning over the last years has led towards a future with autonomous vehicles on our roads. Nevertheless, the performance of their perception systems is strongly dependent on the quality of the utilized training data. As these usually only cover a fraction of all object classes an autonomous driving system will face, such systems struggle with handling the unexpected. In order to safely operate on public roads, the identification of objects from unknown classes remains a crucial task. In this paper, we propose a novel pipeline to detect unknown objects. Instead of focusing on a single sensor modality, we make use of lidar and camera data by combining state-of-the art detection models in a sequential manner. We evaluate our approach on the Waymo Open Perception Dataset and point out current research gaps in anomaly detection.

Neural Language Taskonomy: Which NLP Tasks are the most Predictive of fMRI Brain Activity?

Several popular Transformer based language models have been found to be successful for text-driven brain encoding. However, existing literature leverages only pretrained text Transformer models and has not explored the efficacy of task-specific learned Transformer representations. In this work, we explore transfer learning from representations learned for ten popular natural language processing tasks (two syntactic and eight semantic) for predicting brain responses from two diverse datasets: Pereira (subjects reading sentences from paragraphs) and Narratives (subjects listening to the spoken stories). Encoding models based on task features are used to predict activity in different regions across the whole brain. Features from coreference resolution, NER, and shallow syntax parsing explain greater variance for the reading activity. On the other hand, for the listening activity, tasks such as paraphrase generation, summarization, and natural language inference show better encoding performance. Experiments across all 10 task representations provide the following cognitive insights: (i) language left hemisphere has higher predictive brain activity versus language right hemisphere, (ii) posterior medial cortex, temporo-parieto-occipital junction, dorsal frontal lobe have higher correlation versus early auditory and auditory association cortex, (iii) syntactic and semantic tasks display a good predictive performance across brain regions for reading and listening stimuli resp.

Data Determines Distributional Robustness in Contrastive Language Image Pre-training (CLIP)

Contrastively trained image-text models such as CLIP, ALIGN, and BASIC have demonstrated unprecedented robustness to multiple challenging natural distribution shifts. Since these image-text models differ from previous training approaches in several ways, an important question is what causes the large robustness gains. We answer this question via a systematic experimental investigation. Concretely, we study five different possible causes for the robustness gains: (i) the training set size, (ii) the training distribution, (iii) language supervision at training time, (iv) language supervision at test time, and (v) the contrastive loss function. Our experiments show that the more diverse training distribution is the main cause for the robustness gains, with the other factors contributing little to no robustness. Beyond our experimental results, we also introduce ImageNet-Captions, a version of ImageNet with original text annotations from Flickr, to enable further controlled experiments of language-image training.

Smooth over-parameterized solvers for non-smooth structured optimization

Non-smooth optimization is a core ingredient of many imaging or machine learning pipelines. Non-smoothness encodes structural constraints on the solutions, such as sparsity, group sparsity, low-rank and sharp edges. It is also the basis for the definition of robust loss functions and scale-free functionals such as square-root Lasso. Standard approaches to deal with non-smoothness leverage either proximal splitting or coordinate descent. These approaches are effective but usually require parameter tuning, preconditioning or some sort of support pruning. In this work, we advocate and study a different route, which operates a non-convex but smooth over-parametrization of the underlying non-smooth optimization problems. This generalizes quadratic variational forms that are at the heart of the popular Iterative Reweighted Least Squares (IRLS). Our main theoretical contribution connects gradient descent on this reformulation to a mirror descent flow with a varying Hessian metric. This analysis is crucial to derive convergence bounds that are dimension-free. This explains the efficiency of the method when using small grid sizes in imaging. Our main algorithmic contribution is to apply the Variable Projection (VarPro) method which defines a new formulation by explicitly minimizing over part of the variables. This leads to a better conditioning of the minimized functional and improves the convergence of simple but very efficient gradient-based methods, for instance quasi-Newton solvers. We exemplify the use of this new solver for the resolution of regularized regression problems for inverse problems and supervised learning, including total variation prior and non-convex regularizers.

Deep Learning in Multimodal Remote Sensing Data Fusion: A Comprehensive Review

With the extremely rapid advances in remote sensing (RS) technology, a great quantity of Earth observation (EO) data featuring considerable and complicated heterogeneity is readily available nowadays, which renders researchers an opportunity to tackle current geoscience applications in a fresh way. With the joint utilization of EO data, much research on multimodal RS data fusion has made tremendous progress in recent years, yet these developed traditional algorithms inevitably meet the performance bottleneck due to the lack of the ability to comprehensively analyse and interpret these strongly heterogeneous data. Hence, this non-negligible limitation further arouses an intense demand for an alternative tool with powerful processing competence. Deep learning (DL), as a cutting-edge technology, has witnessed remarkable breakthroughs in numerous computer vision tasks owing to its impressive ability in data representation and reconstruction. Naturally, it has been successfully applied to the field of multimodal RS data fusion, yielding great improvement compared with traditional methods. This survey aims to present a systematic overview in DL-based multimodal RS data fusion. More specifically, some essential knowledge about this topic is first given. Subsequently, a literature survey is conducted to analyse the trends of this field. Some prevalent sub-fields in the multimodal RS data fusion are then reviewed in terms of the to-be-fused data modalities, i.e., spatiospectral, spatiotemporal, light detection and ranging-optical, synthetic aperture radar-optical, and RS-Geospatial Big Data fusion. Furthermore, We collect and summarize some valuable resources for the sake of the development in multimodal RS data fusion. Finally, the remaining challenges and potential future directions are highlighted.

Finding patterns in Knowledge Attribution for Transformers

We analyze the Knowledge Neurons framework for the attribution of factual and relational knowledge to particular neurons in the transformer network. We use a 12-layer multi-lingual BERT model for our experiments. Our study reveals various interesting phenomena. We observe that mostly factual knowledge can be attributed to middle and higher layers of the network($\ge 6$). Further analysis reveals that the middle layers($6-9$) are mostly responsible for relational information, which is further refined into actual factual knowledge or the "correct answer" in the last few layers($10-12$). Our experiments also show that the model handles prompts in different languages, but representing the same fact, similarly, providing further evidence for effectiveness of multi-lingual pre-training. Applying the attribution scheme for grammatical knowledge, we find that grammatical knowledge is far more dispersed among the neurons than factual knowledge.

TracInAD: Measuring Influence for Anomaly Detection

As with many other tasks, neural networks prove very effective for anomaly detection purposes. However, very few deep-learning models are suited for detecting anomalies on tabular datasets. This paper proposes a novel methodology to flag anomalies based on TracIn, an influence measure initially introduced for explicability purposes. The proposed methods can serve to augment any unsupervised deep anomaly detection method. We test our approach using Variational Autoencoders and show that the average influence of a subsample of training points on a test point can serve as a proxy for abnormality. Our model proves to be competitive in comparison with state-of-the-art approaches: it achieves comparable or better performance in terms of detection accuracy on medical and cyber-security tabular benchmark data.

Understanding Urban Water Consumption using Remotely Sensed Data

Urban metabolism is an active field of research that deals with the estimation of emissions and resource consumption from urban regions. The analysis could be carried out through a manual surveyor by the implementation of elegant machine learning algorithms. In this exploratory work, we estimate the water consumption by the buildings in the region captured by satellite imagery. To this end, we break our analysis into three parts: i) Identification of building pixels, given a satellite image, followed by ii) identification of the building type (residential/non-residential) from the building pixels, and finally iii) using the building pixels along with their type to estimate the water consumption using the average per unit area consumption for different building types as obtained from municipal surveys.

Learning Label Initialization for Time-Dependent Harmonic Extension

Node classification on graphs can be formulated as the Dirichlet problem on graphs where the signal is given at the labeled nodes, and the harmonic extension is done on the unlabeled nodes. This paper considers a time-dependent version of the Dirichlet problem on graphs and shows how to improve its solution by learning the proper initialization vector on the unlabeled nodes. Further, we show that the improved solution is at par with state-of-the-art methods used for node classification. Finally, we conclude this paper by discussing the importance of parameter t, pros, and future directions.

Predicting Loose-Fitting Garment Deformations Using Bone-Driven Motion Networks

We present a learning algorithm that uses bone-driven motion networks to predict the deformation of loose-fitting garment meshes at interactive rates. Given a garment, we generate a simulation database and extract virtual bones from simulated mesh sequences using skin decomposition. At runtime, we separately compute low- and high-frequency deformations in a sequential manner. The low-frequency deformations are predicted by transferring body motions to virtual bones' motions, and the high-frequency deformations are estimated leveraging the global information of virtual bones' motions and local information extracted from low-frequency meshes. In addition, our method can estimate garment deformations caused by variations of the simulation parameters (e.g., fabric's bending stiffness) using an RBF kernel ensembling trained networks for different sets of simulation parameters. Through extensive comparisons, we show that our method outperforms state-of-the-art methods in terms of prediction accuracy of mesh deformations by about 20% in RMSE and 10% in Hausdorff distance and STED. The code and data are available at https://github.com/non-void/VirtualBones.

Predicting Issue Types with seBERT

Pre-trained transformer models are the current state-of-the-art for natural language models processing. seBERT is such a model, that was developed based on the BERT architecture, but trained from scratch with software engineering data. We fine-tuned this model for the NLBSE challenge for the task of issue type prediction. Our model dominates the baseline fastText for all three issue types in both recall and precisio} to achieve an overall F1-score of 85.7%, which is an increase of 4.1% over the baseline.

Improved Orientation Estimation and Detection with Hybrid Object Detection Networks for Automotive Radar

This paper presents novel hybrid architectures that combine grid- and point-based processing to improve the detection performance and orientation estimation of radar-based object detection networks. Purely grid-based detection models operate on a bird's-eye-view (BEV) projection of the input point cloud. These approaches suffer from a loss of detailed information through the discrete grid resolution. This applies in particular to radar object detection, where relatively coarse grid resolutions are commonly used to account for the sparsity of radar point clouds. In contrast, point-based models are not affected by this problem as they continuously process point clouds. However, they generally exhibit worse detection performances than grid-based methods. We show that a point-based model can extract neighborhood features, leveraging the exact relative positions of points, before grid rendering. This has significant benefits for a following convolutional detection backbone. In experiments on the public nuScenes dataset our hybrid architecture achieves improvements in terms of detection performance and orientation estimates over networks from previous literature.

Learning Discrete Structured Variational Auto-Encoder using Natural Evolution Strategies

Discrete variational auto-encoders (VAEs) are able to represent semantic latent spaces in generative learning. In many real-life settings, the discrete latent space consists of high-dimensional structures, and propagating gradients through the relevant structures often requires enumerating over an exponentially large latent space. Recently, various approaches were devised to propagate approximated gradients without enumerating over the space of possible structures. In this work, we use Natural Evolution Strategies (NES), a class of gradient-free black-box optimization algorithms, to learn discrete structured VAEs. The NES algorithms are computationally appealing as they estimate gradients with forward pass evaluations only, thus they do not require to propagate gradients through their discrete structures. We demonstrate empirically that optimizing discrete structured VAEs using NES is as effective as gradient-based approximations. Lastly, we prove NES converges for non-Lipschitz functions as appear in discrete structured VAEs.

Open vs Closed-ended questions in attitudinal surveys -- comparing, combining, and interpreting using natural language processing

To improve the traveling experience, researchers have been analyzing the role of attitudes in travel behavior modeling. Although most researchers use closed-ended surveys, the appropriate method to measure attitudes is debatable. Topic Modeling could significantly reduce the time to extract information from open-ended responses and eliminate subjective bias, thereby alleviating analyst concerns. Our research uses Topic Modeling to extract information from open-ended questions and compare its performance with closed-ended responses. Furthermore, some respondents might prefer answering questions using their preferred questionnaire type. So, we propose a modeling framework that allows respondents to use their preferred questionnaire type to answer the survey and enable analysts to use the modeling frameworks of their choice to predict behavior. We demonstrate this using a dataset collected from the USA that measures the intention to use Autonomous Vehicles for commute trips. Respondents were presented with alternative questionnaire versions (open- and closed- ended). Since our objective was also to compare the performance of alternative questionnaire versions, the survey was designed to eliminate influences resulting from statements, behavioral framework, and the choice experiment. Results indicate the suitability of using Topic Modeling to extract information from open-ended responses; however, the models estimated using the closed-ended questions perform better compared to them. Besides, the proposed model performs better compared to the models used currently. Furthermore, our proposed framework will allow respondents to choose the questionnaire type to answer, which could be particularly beneficial to them when using voice-based surveys.

Distilling Governing Laws and Source Input for Dynamical Systems from Videos

Distilling interpretable physical laws from videos has led to expanded interest in the computer vision community recently thanks to the advances in deep learning, but still remains a great challenge. This paper introduces an end-to-end unsupervised deep learning framework to uncover the explicit governing equations of dynamics presented by moving object(s), based on recorded videos. Instead in the pixel (spatial) coordinate system of image space, the physical law is modeled in a regressed underlying physical coordinate system where the physical states follow potential explicit governing equations. A numerical integrator-based sparse regression module is designed and serves as a physical constraint to the autoencoder and coordinate system regression, and, in the meanwhile, uncover the parsimonious closed-form governing equations from the learned physical states. Experiments on simulated dynamical scenes show that the proposed method is able to distill closed-form governing equations and simultaneously identify unknown excitation input for several dynamical systems recorded by videos, which fills in the gap in literature where no existing methods are available and applicable for solving this type of problem.

Disentangled and Side-aware Unsupervised Domain Adaptation for Cross-dataset Subjective Tinnitus Diagnosis

EEG-based tinnitus classification is a valuable tool for tinnitus diagnosis, research, and treatments. Most current works are limited to a single dataset where data patterns are similar. But EEG signals are highly non-stationary, resulting in model's poor generalization to new users, sessions or datasets. Thus, designing a model that can generalize to new datasets is beneficial and indispensable. To mitigate distribution discrepancy across datasets, we propose to achieve Disentangled and Side-aware Unsupervised Domain Adaptation (DSUDA) for cross-dataset tinnitus diagnosis. A disentangled auto-encoder is developed to decouple class-irrelevant information from the EEG signals to improve the classifying ability. The side-aware unsupervised domain adaptation module adapts the class-irrelevant information as domain variance to a new dataset and excludes the variance to obtain the class-distill features for the new dataset classification. It also align signals of left and right ears to overcome inherent EEG pattern difference. We compare DSUDA with state-of-the-art methods, and our model achieves significant improvements over competitors regarding comprehensive evaluation criteria. The results demonstrate our model can successfully generalize to a new dataset and effectively diagnose tinnitus.

FedRN: Exploiting k-Reliable Neighbors Towards Robust Federated Learning

Robustness is becoming another important challenge of federated learning in that the data collection process in each client is naturally accompanied by noisy labels. However, it is far more complex and challenging owing to varying levels of data heterogeneity and noise over clients, which exacerbates the client-to-client performance discrepancy. In this work, we propose a robust federated learning method called FedRN, which exploits k-reliable neighbors with high data expertise or similarity. Our method helps mitigate the gap between low- and high-performance clients by training only with a selected set of clean examples, identified by their ensembled mixture models. We demonstrate the superiority of FedRN via extensive evaluations on three real-world or synthetic benchmark datasets. Compared with existing robust training methods, the results show that FedRN significantly improves the test accuracy in the presence of noisy labels.

CANShield: Signal-based Intrusion Detection for Controller Area Networks

Modern vehicles rely on a fleet of electronic control units (ECUs) connected through controller area network (CAN) buses for critical vehicular control. However, with the expansion of advanced connectivity features in automobiles and the elevated risks of internal system exposure, the CAN bus is increasingly prone to intrusions and injection attacks. The ordinary injection attacks disrupt the typical timing properties of the CAN data stream, and the rule-based intrusion detection systems (IDS) can easily detect them. However, advanced attackers can inject false data to the time series sensory data (signal), while looking innocuous by the pattern/frequency of the CAN messages. Such attacks can bypass the rule-based IDS or any anomaly-based IDS built on binary payload data. To make the vehicles robust against such intelligent attacks, we propose CANShield, a signal-based intrusion detection framework for the CAN bus. CANShield consists of three modules: a data preprocessing module that handles the high-dimensional CAN data stream at the signal level and makes them suitable for a deep learning model; a data analyzer module consisting of multiple deep autoencoder (AE) networks, each analyzing the time-series data from a different temporal perspective; and finally an attack detection module that uses an ensemble method to make the final decision. Evaluation results on two high-fidelity signal-based CAN attack datasets show the high accuracy and responsiveness of CANShield in detecting wide-range of advanced intrusion attacks.

Convergence of Stochastic Approximation via Martingale and Converse Lyapunov Methods

This paper is dedicated to Prof. Eduardo Sontag on the occasion of his seventieth birthday. In this paper, we build upon the ideas first proposed in Gladyshev (1965) to develop a very general framework for proving the almost sure boundedness and the convergence of stochastic approximation algorithms. These ideas are based on martingale methods and are in some ways simpler than convergence proofs based on the ODE method, e.g., Borkar-Meyn (2000). First we study the original version of the SA algorithm introduced in Robbins-Monro (1951), where the objective is to determine a zero of a function, when only noisy measurements of the function are available. The proof makes use of the general framework developed here, together with a new theorem on converse Lyapunov stability, which might be of independent interest. Next we study an alternate version of SA, first introduced in Kiefer-Wolfowitz (1952). The objective here is to find a stationary point of a scalar-valued function, using first-order differences to approximate its gradient. This problem is analyzed in Blum (1954), but with a very opaque proof. We reproduce Blum's conclusions using the proposed framework.

Towards an Ensemble Regressor Model for Anomalous ISP Traffic Prediction

Prediction of network traffic behavior is significant for the effective management of modern telecommunication networks. However, the intuitive approach of predicting network traffic using administrative experience and market analysis data is inadequate for an efficient forecast framework. As a result, many different mathematical models have been studied to capture the general trend of the network traffic and predict accordingly. But the comprehensive performance analysis of varying regression models and their ensemble has not been studied before for analyzing real-world anomalous traffic. In this paper, several regression models such as Extra Gradient Boost (XGBoost), Light Gradient Boosting Machine (LightGBM), Stochastic Gradient Descent (SGD), Gradient Boosting Regressor (GBR), and CatBoost Regressor were analyzed to predict real traffic without and with outliers and show the significance of outlier detection in real-world traffic prediction. Also, we showed the outperformance of the ensemble regression model over the individual prediction model. We compared the performance of different regression models based on five different feature sets of lengths 6, 9, 12, 15, and 18. Our ensemble regression model achieved the minimum average gap of 5.04% between actual and predicted traffic with nine outlier-adjusted inputs. In general, our experimental results indicate that the outliers in the data can significantly impact the quality of the prediction. Thus, outlier detection and mitigation assist the regression model in learning the general trend and making better predictions.

Side-aware Meta-Learning for Cross-Dataset Listener Diagnosis with Subjective Tinnitus

With the development of digital technology, machine learning has paved the way for the next generation of tinnitus diagnoses. Although machine learning has been widely applied in EEG-based tinnitus analysis, most current models are dataset-specific. Each dataset may be limited to a specific range of symptoms, overall disease severity, and demographic attributes; further, dataset formats may differ, impacting model performance. This paper proposes a side-aware meta-learning for cross-dataset tinnitus diagnosis, which can effectively classify tinnitus in subjects of divergent ages and genders from different data collection processes. Owing to the superiority of meta-learning, our method does not rely on large-scale datasets like conventional deep learning models. Moreover, we design a subject-specific training process to assist the model in fitting the data pattern of different patients or healthy people. Our method achieves a high accuracy of 73.8\% in the cross-dataset classification. We conduct an extensive analysis to show the effectiveness of side information of ears in enhancing model performance and side-aware meta-learning in improving the quality of the learned features.

From {Solution Synthesis} to {Student Attempt Synthesis} for Block-Based Visual Programming Tasks

Block-based visual programming environments are increasingly used to introduce computing concepts to beginners. Given that programming tasks are open-ended and conceptual, novice students often struggle when learning in these environments. AI-driven programming tutors hold great promise in automatically assisting struggling students, and need several components to realize this potential. We investigate the crucial component of student modeling, in particular, the ability to automatically infer students' misconceptions for predicting (synthesizing) their behavior. We introduce a novel benchmark, StudentSyn, centered around the following challenge: For a given student, synthesize the student's attempt on a new target task after observing the student's attempt on a fixed reference task. This challenge is akin to that of program synthesis; however, instead of synthesizing a {solution} (i.e., program an expert would write), the goal here is to synthesize a {student attempt} (i.e., program that a given student would write). We first show that human experts (TutorSS) can achieve high performance on the benchmark, whereas simple baselines perform poorly. Then, we develop two neuro/symbolic techniques (NeurSS and SymSS) in a quest to close this gap with TutorSS.

Norm-Agnostic Linear Bandits

Linear bandits have a wide variety of applications including recommendation systems yet they make one strong assumption: the algorithms must know an upper bound $S$ on the norm of the unknown parameter $\theta^*$ that governs the reward generation. Such an assumption forces the practitioner to guess $S$ involved in the confidence bound, leaving no choice but to wish that $\|\theta^*\|\le S$ is true to guarantee that the regret will be low. In this paper, we propose novel algorithms that do not require such knowledge for the first time. Specifically, we propose two algorithms and analyze their regret bounds: one for the changing arm set setting and the other for the fixed arm set setting. Our regret bound for the former shows that the price of not knowing $S$ does not affect the leading term in the regret bound and inflates only the lower order term. For the latter, we do not pay any price in the regret for now knowing $S$. Our numerical experiments show standard algorithms assuming knowledge of $S$ can fail catastrophically when $\|\theta^*\|\le S$ is not true whereas our algorithms enjoy low regret.

Scheduling with Speed Predictions

Algorithms with predictions is a recent framework that has been used to overcome pessimistic worst-case bounds in incomplete information settings. In the context of scheduling, very recent work has leveraged machine-learned predictions to design algorithms that achieve improved approximation ratios in settings where the processing times of the jobs are initially unknown. In this paper, we study the speed-robust scheduling problem where the speeds of the machines, instead of the processing times of the jobs, are unknown and augment this problem with predictions. Our main result is an algorithm that achieves a $\min\{\eta^2(1+\epsilon)^2(1+\alpha), (1+\epsilon)(2 + 2/\alpha)\}$ approximation, for any constants $\alpha, \epsilon \in (0,1)$, where $\eta \geq 1$ is the prediction error. When the predictions are accurate, this approximation improves over the previously best known approximation of $2-1/m$ for speed-robust scheduling, where $m$ is the number of machines, while simultaneously maintaining a worst-case approximation of $(1+\epsilon)(2 + 2/\alpha)$ even when the predictions are wrong. In addition, we obtain improved approximations for the special cases of equal and infinitesimal job sizes, and we complement our algorithmic results with lower bounds. Finally, we empirically evaluate our algorithm against existing algorithms for speed-robust scheduling.

ASTROMER: A transformer-based embedding for the representation of light curves

Taking inspiration from natural language embeddings, we present ASTROMER, a transformer-based model to create representations of light curves. ASTROMER was trained on millions of MACHO R-band samples, and it can be easily fine-tuned to match specific domains associated with downstream tasks. As an example, this paper shows the benefits of using pre-trained representations to classify variable stars. In addition, we provide a python library including all functionalities employed in this work. Our library includes the pre-trained models that can be used to enhance the performance of deep learning models, decreasing computational resources while achieving state-of-the-art results.

Triangular Dropout: Variable Network Width without Retraining

One of the most fundamental design choices in neural networks is layer width: it affects the capacity of what a network can learn and determines the complexity of the solution. This latter property is often exploited when introducing information bottlenecks, forcing a network to learn compressed representations. However, such an architecture decision is typically immutable once training begins; switching to a more compressed architecture requires retraining. In this paper we present a new layer design, called Triangular Dropout, which does not have this limitation. After training, the layer can be arbitrarily reduced in width to exchange performance for narrowness. We demonstrate the construction and potential use cases of such a mechanism in three areas. Firstly, we describe the formulation of Triangular Dropout in autoencoders, creating models with selectable compression after training. Secondly, we add Triangular Dropout to VGG19 on ImageNet, creating a powerful network which, without retraining, can be significantly reduced in parameters. Lastly, we explore the application of Triangular Dropout to reinforcement learning (RL) policies on selected control problems.

The Limits of Word Level Differential Privacy

As the issues of privacy and trust are receiving increasing attention within the research community, various attempts have been made to anonymize textual data. A significant subset of these approaches incorporate differentially private mechanisms to perturb word embeddings, thus replacing individual words in a sentence. While these methods represent very important contributions, have various advantages over other techniques and do show anonymization capabilities, they have several shortcomings. In this paper, we investigate these weaknesses and demonstrate significant mathematical constraints diminishing the theoretical privacy guarantee as well as major practical shortcomings with regard to the protection against deanonymization attacks, the preservation of content of the original sentences as well as the quality of the language output. Finally, we propose a new method for text anonymization based on transformer based language models fine-tuned for paraphrasing that circumvents most of the identified weaknesses and also offers a formal privacy guarantee. We evaluate the performance of our method via thorough experimentation and demonstrate superior performance over the discussed mechanisms.

One Weird Trick to Improve Your Semi-Weakly Supervised Semantic Segmentation Model

Semi-weakly supervised semantic segmentation (SWSSS) aims to train a model to identify objects in images based on a small number of images with pixel-level labels, and many more images with only image-level labels. Most existing SWSSS algorithms extract pixel-level pseudo-labels from an image classifier - a very difficult task to do well, hence requiring complicated architectures and extensive hyperparameter tuning on fully-supervised validation sets. We propose a method called prediction filtering, which instead of extracting pseudo-labels, just uses the classifier as a classifier: it ignores any segmentation predictions from classes which the classifier is confident are not present. Adding this simple post-processing method to baselines gives results competitive with or better than prior SWSSS algorithms. Moreover, it is compatible with pseudo-label methods: adding prediction filtering to existing SWSSS algorithms further improves segmentation performance.

Retrieval-Enhanced Machine Learning

Although information access systems have long supported people in accomplishing a wide range of tasks, we propose broadening the scope of users of information access systems to include task-driven machines, such as machine learning models. In this way, the core principles of indexing, representation, retrieval, and ranking can be applied and extended to substantially improve model generalization, scalability, robustness, and interpretability. We describe a generic retrieval-enhanced machine learning (REML) framework, which includes a number of existing models as special cases. REML challenges information retrieval conventions, presenting opportunities for novel advances in core areas, including optimization. The REML research agenda lays a foundation for a new style of information access research and paves a path towards advancing machine learning and artificial intelligence.

COMET Flows: Towards Generative Modeling of Multivariate Extremes and Tail Dependence

Normalizing flows, a popular class of deep generative models, often fail to represent extreme phenomena observed in real-world processes. In particular, existing normalizing flow architectures struggle to model multivariate extremes, characterized by heavy-tailed marginal distributions and asymmetric tail dependence among variables. In light of this shortcoming, we propose COMET (COpula Multivariate ExTreme) Flows, which decompose the process of modeling a joint distribution into two parts: (i) modeling its marginal distributions, and (ii) modeling its copula distribution. COMET Flows capture heavy-tailed marginal distributions by combining a parametric tail belief at extreme quantiles of the marginals with an empirical kernel density function at mid-quantiles. In addition, COMET Flows capture asymmetric tail dependence among multivariate extremes by viewing such dependence as inducing a low-dimensional manifold structure in feature space. Experimental results on both synthetic and real-world datasets demonstrate the effectiveness of COMET Flows in capturing both heavy-tailed marginals and asymmetric tail dependence compared to other state-of-the-art baseline architectures. All code is available on GitHub at https://github.com/andrewmcdonald27/COMETFlows.

FINETUNA: Fine-tuning Accelerated Molecular Simulations

Machine learning approaches have the potential to approximate Density Functional Theory (DFT) for atomistic simulations in a computationally efficient manner, which could dramatically increase the impact of computational simulations on real-world problems. However, they are limited by their accuracy and the cost of generating labeled data. Here, we present an online active learning framework for accelerating the simulation of atomic systems efficiently and accurately by incorporating prior physical information learned by large-scale pre-trained graph neural network models from the Open Catalyst Project. Accelerating these simulations enables useful data to be generated more cheaply, allowing better models to be trained and more atomistic systems to be screened. We also present a method of comparing local optimization techniques on the basis of both their speed and accuracy. Experiments on 30 benchmark adsorbate-catalyst systems show that our method of transfer learning to incorporate prior information from pre-trained models accelerates simulations by reducing the number of DFT calculations by 91%, while meeting an accuracy threshold of 0.02 eV 93% of the time. Finally, we demonstrate a technique for leveraging the interactive functionality built in to VASP to efficiently compute single point calculations within our online active learning framework without the significant startup costs. This allows VASP to work in tandem with our framework while requiring 75% fewer self-consistent cycles than conventional single point calculations. The online active learning implementation, and examples using the VASP interactive code, are available in the open source FINETUNA package on Github.

Leveraging Stochastic Predictions of Bayesian Neural Networks for Fluid Simulations

We investigate uncertainty estimation and multimodality via the non-deterministic predictions of Bayesian neural networks (BNNs) in fluid simulations. To this end, we deploy BNNs in three challenging experimental test-cases of increasing complexity: We show that BNNs, when used as surrogate models for steady-state fluid flow predictions, provide accurate physical predictions together with sensible estimates of uncertainty. Further, we experiment with perturbed temporal sequences from Navier-Stokes simulations and evaluate the capabilities of BNNs to capture multimodal evolutions. While our findings indicate that this is problematic for large perturbations, our results show that the networks learn to correctly predict high uncertainties in such situations. Finally, we study BNNs in the context of solver interactions with turbulent plasma flows. We find that BNN-based corrector networks can stabilize coarse-grained simulations and successfully create multimodal trajectories.

An improvement to a result about graph isomorphism networks using the prime factorization theorem

The unique prime factorization theorem is used to show the existence of a function on a countable set $\mathcal{X}$ so that the sum aggregator function is injective on all multisets of $\mathcal{X}$ of finite size.

Streaming Inference for Infinite Non-Stationary Clustering

Learning from a continuous stream of non-stationary data in an unsupervised manner is arguably one of the most common and most challenging settings facing intelligent agents. Here, we attack learning under all three conditions (unsupervised, streaming, non-stationary) in the context of clustering, also known as mixture modeling. We introduce a novel clustering algorithm that endows mixture models with the ability to create new clusters online, as demanded by the data, in a probabilistic, time-varying, and principled manner. To achieve this, we first define a novel stochastic process called the Dynamical Chinese Restaurant Process (Dynamical CRP), which is a non-exchangeable distribution over partitions of a set; next, we show that the Dynamical CRP provides a non-stationary prior over cluster assignments and yields an efficient streaming variational inference algorithm. We conclude with experiments showing that the Dynamical CRP can be applied on diverse synthetic and real data with Gaussian and non-Gaussian likelihoods.

Applications of Deep Learning to the Design of Enhanced Wireless Communication Systems

Innovation in the physical layer of communication systems has traditionally been achieved by breaking down the transceivers into sets of processing blocks, each optimized independently based on mathematical models. Conversely, deep learning (DL)-based systems are able to handle increasingly complex tasks for which no tractable models are available. This thesis aims at comparing different approaches to unlock the full potential of DL in the physical layer. First, we describe a neural network (NN)-based block strategy, where an NN is optimized to replace a block in a communication system. We apply this strategy to introduce a multi-user multiple-input multiple-output (MU-MIMO) detector that builds on top of an existing DL-based architecture. Second, we detail an end-to-end strategy, in which the transmitter and receiver are modeled as an autoencoder. This approach is illustrated with the design of waveforms that achieve high throughputs while satisfying peak-to-average power ratio (PAPR) and adjacent channel leakage ratio (ACLR) constraints. Lastly, we propose a hybrid strategy, where multiple DL components are inserted into a traditional architecture but are trained to optimize the end-to-end performance. To demonstrate its benefits, we propose a DL-enhanced MU-MIMO receiver that both enable lower bit error rates (BERs) compared to a conventional receiver and remains scalable to any number of users. Each approach has its own strengths and shortcomings. While the first one is the easiest to implement, its individual block optimization does not ensure the overall system optimality. On the other hand, systems designed with the second approach are computationally complex but allow for new opportunities such as pilotless transmissions. Finally, the combined flexibility and end-to-end performance gains of the third approach motivate its use for short-term practical implementations.

Multi-Task Text Classification using Graph Convolutional Networks for Large-Scale Low Resource Language

Graph Convolutional Networks (GCN) have achieved state-of-art results on single text classification tasks like sentiment analysis, emotion detection, etc. However, the performance is achieved by testing and reporting on resource-rich languages like English. Applying GCN for multi-task text classification is an unexplored area. Moreover, training a GCN or adopting an English GCN for Indian languages is often limited by data availability, rich morphological variation, syntax, and semantic differences. In this paper, we study the use of GCN for the Telugu language in single and multi-task settings for four natural language processing (NLP) tasks, viz. sentiment analysis (SA), emotion identification (EI), hate-speech (HS), and sarcasm detection (SAR). In order to evaluate the performance of GCN with one of the Indian languages, Telugu, we analyze the GCN based models with extensive experiments on four downstream tasks. In addition, we created an annotated Telugu dataset, TEL-NLP, for the four NLP tasks. Further, we propose a supervised graph reconstruction method, Multi-Task Text GCN (MT-Text GCN) on the Telugu that leverages to simultaneously (i) learn the low-dimensional word and sentence graph embeddings from word-sentence graph reconstruction using graph autoencoder (GAE) and (ii) perform multi-task text classification using these latent sentence graph embeddings. We argue that our proposed MT-Text GCN achieves significant improvements on TEL-NLP over existing Telugu pretrained word embeddings, and multilingual pretrained Transformer models: mBERT, and XLM-R. On TEL-NLP, we achieve a high F1-score for four NLP tasks: SA (0.84), EI (0.55), HS (0.83) and SAR (0.66). Finally, we show our model's quantitative and qualitative analysis on the four NLP tasks in Telugu.

Predicting Time-to-conversion for Dementia of Alzheimer's Type using Multi-modal Deep Survival Analysis

Dementia of Alzheimer's Type (DAT) is a complex disorder influenced by numerous factors, but it is unclear how each factor contributes to disease progression. An in-depth examination of these factors may yield an accurate estimate of time-to-conversion to DAT for patients at various disease stages. We used 401 subjects with 63 features from MRI, genetic, and CDC (Cognitive tests, Demographic, and CSF) data modalities in the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. We used a deep learning-based survival analysis model that extends the classic Cox regression model to predict time-to-conversion to DAT. Our findings showed that genetic features contributed the least to survival analysis, while CDC features contributed the most. Combining MRI and genetic features improved survival prediction over using either modality alone, but adding CDC to any combination of features only worked as well as using only CDC features. Consequently, our study demonstrated that using the current clinical procedure, which includes gathering cognitive test results, can outperform survival analysis results produced using costly genetic or CSF data.

Performance Weighting for Robust Federated Learning Against Corrupted Sources

Federated Learning has emerged as a dominant computational paradigm for distributed machine learning. Its unique data privacy properties allow us to collaboratively train models while offering participating clients certain privacy-preserving guarantees. However, in real-world applications, a federated environment may consist of a mixture of benevolent and malicious clients, with the latter aiming to corrupt and degrade federated model's performance. Different corruption schemes may be applied such as model poisoning and data corruption. Here, we focus on the latter, the susceptibility of federated learning to various data corruption attacks. We show that the standard global aggregation scheme of local weights is inefficient in the presence of corrupted clients. To mitigate this problem, we propose a class of task-oriented performance-based methods computed over a distributed validation dataset with the goal to detect and mitigate corrupted clients. Specifically, we construct a robust weight aggregation scheme based on geometric mean and demonstrate its effectiveness under random label shuffling and targeted label flipping attacks.

Using Machine Learning to Evaluate Real Estate Prices Using Location Big Data

With everyone trying to enter the real estate market nowadays, knowing the proper valuations for residential and commercial properties has become crucial. Past researchers have been known to utilize static real estate data (e.g. number of beds, baths, square footage) or even a combination of real estate and demographic information to predict property prices. In this investigation, we attempted to improve upon past research. So we decided to explore a unique approach: we wanted to determine if mobile location data could be used to improve the predictive power of popular regression and tree-based models. To prepare our data for our models, we processed the mobility data by attaching it to individual properties from the real estate data that aggregated users within 500 meters of the property for each day of the week. We removed people that lived within 500 meters of each property, so each property's aggregated mobility data only contained non-resident census features. On top of these dynamic census features, we also included static census features, including the number of people in the area, the average proportion of people commuting, and the number of residents in the area. Finally, we tested multiple models to predict real estate prices. Our proposed model is two stacked random forest modules combined using a ridge regression that uses the random forest outputs as predictors. The first random forest model used static features only and the second random forest model used dynamic features only. Comparing our models with and without the dynamic mobile location features concludes the model with dynamic mobile location features achieves 3/% percent lower mean squared error than the same model but without dynamic mobile location features.

VAE-Loco: Versatile Quadruped Locomotion by Learning a Disentangled Gait Representation

Quadruped locomotion is rapidly maturing to a degree where robots now routinely traverse a variety of unstructured terrains. However, while gaits can be varied typically by selecting from a range of pre-computed styles, current planners are unable to vary key gait parameters continuously while the robot is in motion. The synthesis, on-the-fly, of gaits with unexpected operational characteristics or even the blending of dynamic manoeuvres lies beyond the capabilities of the current state-of-the-art. In this work we address this limitation by learning a latent space capturing the key stance phases constituting a particular gait. This is achieved via a generative model trained on a single trot style, which encourages disentanglement such that application of a drive signal to a single dimension of the latent state induces holistic plans synthesising a continuous variety of trot styles. We demonstrate that specific properties of the drive signal map directly to gait parameters such as cadence, footstep height and full stance duration. Due to the nature of our approach these synthesised gaits are continuously variable online during robot operation and robustly capture a richness of movement significantly exceeding the relatively narrow behaviour seen during training. In addition, the use of a generative model facilitates the detection and mitigation of disturbances to provide a versatile and robust planning framework. We evaluate our approach on two versions of the real ANYmal quadruped robots and demonstrate that our method achieves a continuous blend of dynamic trot styles whilst being robust and reactive to external perturbations.

Reproducing Kernels and New Approaches in Compositional Data Analysis

Compositional data, such as human gut microbiomes, consist of non-negative variables whose only the relative values to other variables are available. Analyzing compositional data such as human gut microbiomes needs a careful treatment of the geometry of the data. A common geometrical understanding of compositional data is via a regular simplex. Majority of existing approaches rely on a log-ratio or power transformations to overcome the innate simplicial geometry. In this work, based on the key observation that a compositional data are projective in nature, and on the intrinsic connection between projective and spherical geometry, we re-interpret the compositional domain as the quotient topology of a sphere modded out by a group action. This re-interpretation allows us to understand the function space on compositional domains in terms of that on spheres and to use spherical harmonics theory along with reflection group actions for constructing a compositional Reproducing Kernel Hilbert Space (RKHS). This construction of RKHS for compositional data will widely open research avenues for future methodology developments. In particular, well-developed kernel embedding methods can be now introduced to compositional data analysis. The polynomial nature of compositional RKHS has both theoretical and computational benefits. The wide applicability of the proposed theoretical framework is exemplified with nonparametric density estimation and kernel exponential family for compositional data.

SELC: Self-Ensemble Label Correction Improves Learning with Noisy Labels

Deep neural networks are prone to overfitting noisy labels, resulting in poor generalization performance. To overcome this problem, we present a simple and effective method self-ensemble label correction (SELC) to progressively correct noisy labels and refine the model. We look deeper into the memorization behavior in training with noisy labels and observe that the network outputs are reliable in the early stage. To retain this reliable knowledge, SELC uses ensemble predictions formed by an exponential moving average of network outputs to update the original noisy labels. We show that training with SELC refines the model by gradually reducing supervision from noisy labels and increasing supervision from ensemble predictions. Despite its simplicity, compared with many state-of-the-art methods, SELC obtains more promising and stable results in the presence of class-conditional, instance-dependent, and real-world label noise. The code is available at https://github.com/MacLLL/SELC.

Emotion-Controllable Generalized Talking Face Generation

Despite the significant progress in recent years, very few of the AI-based talking face generation methods attempt to render natural emotions. Moreover, the scope of the methods is majorly limited to the characteristics of the training dataset, hence they fail to generalize to arbitrary unseen faces. In this paper, we propose a one-shot facial geometry-aware emotional talking face generation method that can generalize to arbitrary faces. We propose a graph convolutional neural network that uses speech content feature, along with an independent emotion input to generate emotion and speech-induced motion on facial geometry-aware landmark representation. This representation is further used in our optical flow-guided texture generation network for producing the texture. We propose a two-branch texture generation network, with motion and texture branches designed to consider the motion and texture content independently. Compared to the previous emotion talking face methods, our method can adapt to arbitrary faces captured in-the-wild by fine-tuning with only a single image of the target identity in neutral emotion.

D-DPCC: Deep Dynamic Point Cloud Compression via 3D Motion Prediction

The non-uniformly distributed nature of the 3D dynamic point cloud (DPC) brings significant challenges to its high-efficient inter-frame compression. This paper proposes a novel 3D sparse convolution-based Deep Dynamic Point Cloud Compression (D-DPCC) network to compensate and compress the DPC geometry with 3D motion estimation and motion compensation in the feature space. In the proposed D-DPCC network, we design a {\it Multi-scale Motion Fusion} (MMF) module to accurately estimate the 3D optical flow between the feature representations of adjacent point cloud frames. Specifically, we utilize a 3D sparse convolution-based encoder to obtain the latent representation for motion estimation in the feature space and introduce the proposed MMF module for fused 3D motion embedding. Besides, for motion compensation, we propose a 3D {\it Adaptively Weighted Interpolation} (3DAWI) algorithm with a penalty coefficient to adaptively decrease the impact of distant neighbors. We compress the motion embedding and the residual with a lossy autoencoder-based network. To our knowledge, this paper is the first work proposing an end-to-end deep dynamic point cloud compression framework. The experimental result shows that the proposed D-DPCC framework achieves an average 76\% BD-Rate (Bjontegaard Delta Rate) gains against state-of-the-art Video-based Point Cloud Compression (V-PCC) v13 in inter mode.

Hausa Visual Genome: A Dataset for Multi-Modal English to Hausa Machine Translation

Multi-modal Machine Translation (MMT) enables the use of visual information to enhance the quality of translations. The visual information can serve as a valuable piece of context information to decrease the ambiguity of input sentences. Despite the increasing popularity of such a technique, good and sizeable datasets are scarce, limiting the full extent of their potential. Hausa, a Chadic language, is a member of the Afro-Asiatic language family. It is estimated that about 100 to 150 million people speak the language, with more than 80 million indigenous speakers. This is more than any of the other Chadic languages. Despite a large number of speakers, the Hausa language is considered low-resource in natural language processing (NLP). This is due to the absence of sufficient resources to implement most NLP tasks. While some datasets exist, they are either scarce, machine-generated, or in the religious domain. Therefore, there is a need to create training and evaluation data for implementing machine learning tasks and bridging the research gap in the language. This work presents the Hausa Visual Genome (HaVG), a dataset that contains the description of an image or a section within the image in Hausa and its equivalent in English. To prepare the dataset, we started by translating the English description of the images in the Hindi Visual Genome (HVG) into Hausa automatically. Afterward, the synthetic Hausa data was carefully post-edited considering the respective images. The dataset comprises 32,923 images and their descriptions that are divided into training, development, test, and challenge test set. The Hausa Visual Genome is the first dataset of its kind and can be used for Hausa-English machine translation, multi-modal research, and image description, among various other natural language processing and generation tasks.

A Deep Learning-based Integrated Framework for Quality-aware Undersampled Cine Cardiac MRI Reconstruction and Analysis

Cine cardiac magnetic resonance (CMR) imaging is considered the gold standard for cardiac function evaluation. However, cine CMR acquisition is inherently slow and in recent decades considerable effort has been put into accelerating scan times without compromising image quality or the accuracy of derived results. In this paper, we present a fully-automated, quality-controlled integrated framework for reconstruction, segmentation and downstream analysis of undersampled cine CMR data. The framework enables active acquisition of radial k-space data, in which acquisition can be stopped as soon as acquired data are sufficient to produce high quality reconstructions and segmentations. This results in reduced scan times and automated analysis, enabling robust and accurate estimation of functional biomarkers. To demonstrate the feasibility of the proposed approach, we perform realistic simulations of radial k-space acquisitions on a dataset of subjects from the UK Biobank and present results on in-vivo cine CMR k-space data collected from healthy subjects. The results demonstrate that our method can produce quality-controlled images in a mean scan time reduced from 12 to 4 seconds per slice, and that image quality is sufficient to allow clinically relevant parameters to be automatically estimated to within 5% mean absolute difference.

ComPhy: Compositional Physical Reasoning of Objects and Events from Videos

Objects' motions in nature are governed by complex interactions and their properties. While some properties, such as shape and material, can be identified via the object's visual appearances, others like mass and electric charge are not directly visible. The compositionality between the visible and hidden properties poses unique challenges for AI models to reason from the physical world, whereas humans can effortlessly infer them with limited observations. Existing studies on video reasoning mainly focus on visually observable elements such as object appearance, movement, and contact interaction. In this paper, we take an initial step to highlight the importance of inferring the hidden physical properties not directly observable from visual appearances, by introducing the Compositional Physical Reasoning (ComPhy) dataset. For a given set of objects, ComPhy includes few videos of them moving and interacting under different initial conditions. The model is evaluated based on its capability to unravel the compositional hidden properties, such as mass and charge, and use this knowledge to answer a set of questions posted on one of the videos. Evaluation results of several state-of-the-art video reasoning models on ComPhy show unsatisfactory performance as they fail to capture these hidden properties. We further propose an oracle neural-symbolic framework named Compositional Physics Learner (CPL), combining visual perception, physical property learning, dynamic prediction, and symbolic execution into a unified framework. CPL can effectively identify objects' physical properties from their interactions and predict their dynamics to answer questions.

Ensemble pruning via an integer programming approach with diversity constraints

Ensemble learning combines multiple classifiers in the hope of obtaining better predictive performance. Empirical studies have shown that ensemble pruning, that is, choosing an appropriate subset of the available classifiers, can lead to comparable or better predictions than using all classifiers. In this paper, we consider a binary classification problem and propose an integer programming (IP) approach for selecting optimal classifier subsets. We propose a flexible objective function to adapt to desired criteria of different datasets. We also propose constraints to ensure minimum diversity levels in the ensemble. Despite the general case of IP being NP-Hard, state-of-the-art solvers are able to quickly obtain good solutions for datasets with up to 60000 data points. Our approach yields competitive results when compared to some of the best and most used pruning methods in literature.