Search

Scalable service-Oriented Middleware over IP (SOME/IP) is an Ethernet communication standard protocol in the Automotive Open System Architecture (AUTOSAR), promoting ECU-to-ECU communication over the IP stack. However, SOME/IP lacks a robust security architecture, making it susceptible to potential attacks. Besides, random hardware failure of ECU will disrupt SOME/IP communication. In this paper, we propose SISSA, a SOME/IP communication traffic-based approach for modeling and analyzing in-vehicle functional safety and cyber security. Specifically, SISSA models hardware failures with the Weibull distribution and addresses five potential attacks on SOME/IP communication, including Distributed Denial-of-Services, Man-in-the-Middle, and abnormal communication processes, assuming a malicious user accesses the in-vehicle network. Subsequently, SISSA designs a series of deep learning models with various backbones to extract features from SOME/IP sessions among ECUs. We adopt residual self-attention to accelerate the model's convergence and enhance detection accuracy, determining whether an ECU is under attack, facing functional failure, or operating normally. Additionally, we have created and annotated a dataset encompassing various classes, including indicators of attack, functionality, and normalcy. This contribution is noteworthy due to the scarcity of publicly accessible datasets with such characteristics.Extensive experimental results show the effectiveness and efficiency of SISSA.

The Fisher information matrix characterizes the local geometry in the parameter space of neural networks. It elucidates insightful theories and useful tools to understand and optimize neural networks. Given its high computational cost, practitioners often use random estimators and evaluate only the diagonal entries. We examine two such estimators, whose accuracy and sample complexity depend on their associated variances. We derive bounds of the variances and instantiate them in regression and classification networks. We navigate trade-offs of both estimators based on analytical and numerical studies. We find that the variance quantities depend on the non-linearity with respect to different parameter groups and should not be neglected when estimating the Fisher information.

Deep models are dominating the artificial intelligence (AI) industry since the ImageNet challenge in 2012. The size of deep models is increasing ever since, which brings new challenges to this field with applications in cell phones, personal computers, autonomous cars, and wireless base stations. Here we list a set of problems, ranging from training, inference, generalization bound, and optimization with some formalism to communicate these challenges with mathematicians, statisticians, and theoretical computer scientists. This is a subjective view of the research questions in deep learning that benefits the tech industry in long run.

We study the problem of imputing missing values in a dataset, which has important applications in many domains. The key to missing value imputation is to capture the data distribution with incomplete samples and impute the missing values accordingly. In this paper, by leveraging the fact that any two batches of data with missing values come from the same data distribution, we propose to impute the missing values of two batches of samples by transforming them into a latent space through deep invertible functions and matching them distributionally. To learn the transformations and impute the missing values simultaneously, a simple and well-motivated algorithm is proposed. Our algorithm has fewer hyperparameters to fine-tune and generates high-quality imputations regardless of how missing values are generated. Extensive experiments over a large number of datasets and competing benchmark algorithms show that our method achieves state-of-the-art performance., Comment: ICML 2023 camera-ready version, https://openreview.net/forum?id=WBWb1FU8iz

In this paper, $2\times2$ zero-sum games are studied under the following assumptions: $(1)$ One of the players (the leader) commits to choose its actions by sampling a given probability measure (strategy); $(2)$ The leader announces its action, which is observed by its opponent (the follower) through a binary channel; and $(3)$ the follower chooses its strategy based on the knowledge of the leader's strategy and the noisy observation of the leader's action. Under these conditions, the equilibrium is shown to always exist. Interestingly, even subject to noise, observing the actions of the leader is shown to be either beneficial or immaterial for the follower. More specifically, the payoff at the equilibrium of this game is upper bounded by the payoff at the Stackelberg equilibrium (SE) in pure strategies; and lower bounded by the payoff at the Nash equilibrium, which is equivalent to the SE in mixed strategies.Finally, necessary and sufficient conditions for observing the payoff at equilibrium to be equal to its lower bound are presented. Sufficient conditions for the payoff at equilibrium to be equal to its upper bound are also presented., Comment: Accepted by 2023 IEEE Int. Symp. on Information Theory (ISIT)

We consider the problem of learning a set of probability distributions from the Bellman dynamics in distributional reinforcement learning~(RL) that learns the whole return distribution compared with only its expectation in classical RL. Despite its success to obtain superior performance, we still have a poor understanding of how the value distribution in distributional RL works. In this study, we analyze the optimization benefits of distributional RL by leverage of additional value distribution information over classical RL in the Neural Fitted Z-Iteration~(Neural FZI) framework. To begin with, we demonstrate that the distribution loss of distributional RL has desirable smoothness characteristics and hence enjoys stable gradients, which is in line with its tendency to promote optimization stability. Furthermore, the acceleration effect of distributional RL is revealed by decomposing the return distribution. It turns out that distributional RL can perform favorably if the value distribution approximation is appropriate, measured by the variance of gradient estimates in each environment for any specific distributional RL algorithm. Rigorous experiments validate the stable optimization behaviors of distributional RL, contributing to its acceleration effects compared to classical RL. The findings of our research illuminate how the value distribution in distributional RL algorithms helps the optimization., Comment: arXiv admin note: text overlap with arXiv:2110.03155

A key technique of machine learning and computer vision is to embed discrete weighted graphs into continuous spaces for further downstream processing. Embedding discrete hierarchical structures in hyperbolic geometry has proven very successful since it was shown that any weighted tree can be embedded in that geometry with arbitrary low distortion. Various optimization methods for hyperbolic embeddings based on common models of hyperbolic geometry have been studied. In this paper, we consider Hilbert geometry for the standard simplex which is isometric to a vector space equipped with the variation polytope norm. We study the representation power of this Hilbert simplex geometry by embedding distance matrices of graphs. Our findings demonstrate that Hilbert simplex geometry is competitive to alternative geometries such as the Poincar\'e hyperbolic ball or the Euclidean geometry for embedding tasks while being fast and numerically robust., Comment: 21 pages, 12 figures

Modeling stochastic traffic dynamics is critical to developing self-driving cars. Because it is difficult to develop first principle models of cars driven by humans, there is great potential for using data driven approaches in developing traffic dynamical models. While there is extensive literature on this subject, previous works mainly address the prediction accuracy of data-driven models. Moreover, it is often difficult to apply these models to common planning frameworks since they fail to meet the assumptions therein. In this work, we propose a new stochastic traffic model, Recurrent Traffic Graph Neural Network (RTGNN), by enforcing additional structures on the model so that the proposed model can be seamlessly integrated with existing motion planning algorithms. RTGNN is a Markovian model and is able to infer future traffic states conditioned on the motion of the ego vehicle. Specifically, RTGNN uses a definition of the traffic state that includes the state of all players in a local region and is therefore able to make joint predictions for all agents of interest. Meanwhile, we explicitly model the hidden states of agents, "intentions," as part of the traffic state to reflect the inherent partial observability of traffic dynamics. The above mentioned properties are critical for integrating RTGNN with motion planning algorithms coupling prediction and decision making. Despite the additional structures, we show that RTGNN is able to achieve state-of-the-art accuracy through comparisons with other similar works., Comment: Accepted by ICRA 2022

The empirical success of distributional reinforcement learning~(RL) highly depends on the distribution representation and the choice of distribution divergence. In this paper, we propose \textit{Sinkhorn distributional RL~(SinkhornDRL)} that learns unrestricted statistics from return distributions and leverages Sinkhorn divergence to minimize the difference between current and target Bellman return distributions. Theoretically, we prove the contraction properties of SinkhornDRL, consistent with the interpolation nature of Sinkhorn divergence between Wasserstein distance and Maximum Mean Discrepancy~(MMD). We also establish the equivalence between Sinkhorn divergence and a regularized MMD with a regularized Moment Matching behavior, contributing to explaining the superiority of SinkhornDRL. Empirically, we show that SinkhornDRL is consistently better or comparable to existing algorithms on the Atari games suite., Comment: arXiv admin note: text overlap with arXiv:2110.03155

We introduce a new family of techniques to post-process ("wrap") a black-box classifier in order to reduce its bias. Our technique builds on the recent analysis of improper loss functions whose optimization can correct any twist in prediction, unfairness being treated as a twist. In the post-processing, we learn a wrapper function which we define as an $\alpha$-tree, which modifies the prediction. We provide two generic boosting algorithms to learn $\alpha$-trees. We show that our modification has appealing properties in terms of composition of $\alpha$-trees, generalization, interpretability, and KL divergence between modified and original predictions. We exemplify the use of our technique in three fairness notions: conditional value-at-risk, equality of opportunity, and statistical parity; and provide experiments on several readily available datasets., Comment: Published in Advances in Neural Information Processing Systems 35 (NeurIPS 2022)

Graph contrastive learning attracts/disperses node representations for similar/dissimilar node pairs under some notion of similarity. It may be combined with a low-dimensional embedding of nodes to preserve intrinsic and structural properties of a graph. In this paper, we extend the celebrated Laplacian Eigenmaps with contrastive learning, and call them COntrastive Laplacian EigenmapS (COLES). Starting from a GAN-inspired contrastive formulation, we show that the Jensen-Shannon divergence underlying many contrastive graph embedding models fails under disjoint positive and negative distributions, which may naturally emerge during sampling in the contrastive setting. In contrast, we demonstrate analytically that COLES essentially minimizes a surrogate of Wasserstein distance, which is known to cope well under disjoint distributions. Moreover, we show that the loss of COLES belongs to the family of so-called block-contrastive losses, previously shown to be superior compared to pair-wise losses typically used by contrastive methods. We show on popular benchmarks/backbones that COLES offers favourable accuracy/scalability compared to DeepWalk, GCN, Graph2Gauss, DGI and GRACE baselines., Comment: Accepted by NeurIPS 2021. Includes the main paper and the supplementary material. OpenReview: https://openreview.net/forum?id=iLn-bhP-kKH

Adversarial robustness, which primarily comprises sensitivity-based robustness and spatial robustness, plays an integral part in achieving robust generalization. In this paper, we endeavor to design strategies to achieve universal adversarial robustness. To achieve this, we first investigate the relatively less-explored realm of spatial robustness. Then, we integrate the existing spatial robustness methods by incorporating both local and global spatial vulnerability into a unified spatial attack and adversarial training approach. Furthermore, we present a comprehensive relationship between natural accuracy, sensitivity-based robustness, and spatial robustness, supported by strong evidence from the perspective of robust representation. Crucially, to reconcile the interplay between the mutual impacts of various robustness components into one unified framework, we incorporate the \textit{Pareto criterion} into the adversarial robustness analysis, yielding a novel strategy called Pareto Adversarial Training for achieving universal robustness. The resulting Pareto front, which delineates the set of optimal solutions, provides an optimal balance between natural accuracy and various adversarial robustness. This sheds light on solutions for achieving universal robustness in the future. To the best of our knowledge, we are the first to consider universal adversarial robustness via multi-objective optimization., Comment: Published in SCIENCE CHINA Information Sciences (SCIS) in 2023. Please also refer to the published version in the Journal reference https://www.sciengine.com/SCIS/doi/10.1007/s11432-022-3861-8

Anderson mixing has been heuristically applied to reinforcement learning (RL) algorithms for accelerating convergence and improving the sampling efficiency of deep RL. Despite its heuristic improvement of convergence, a rigorous mathematical justification for the benefits of Anderson mixing in RL has not yet been put forward. In this paper, we provide deeper insights into a class of acceleration schemes built on Anderson mixing that improve the convergence of deep RL algorithms. Our main results establish a connection between Anderson mixing and quasi-Newton methods and prove that Anderson mixing increases the convergence radius of policy iteration schemes by an extra contraction factor. The key focus of the analysis roots in the fixed-point iteration nature of RL. We further propose a stabilization strategy by introducing a stable regularization term in Anderson mixing and a differentiable, non-expansive MellowMax operator that can allow both faster convergence and more stable behavior. Extensive experiments demonstrate that our proposed method enhances the convergence, stability, and performance of RL algorithms.

We aim at capturing high-order statistics of feature vectors formed by a neural network, and propose end-to-end second- and higher-order pooling to form a tensor descriptor. Tensor descriptors require a robust similarity measure due to low numbers of aggregated vectors and the burstiness phenomenon, when a given feature appears more/less frequently than statistically expected. The Heat Diffusion Process (HDP) on a graph Laplacian is closely related to the Eigenvalue Power Normalization (EPN) of the covariance/autocorrelation matrix, whose inverse forms a loopy graph Laplacian. We show that the HDP and the EPN play the same role, i.e., to boost or dampen the magnitude of the eigenspectrum thus preventing the burstiness. We equip higher-order tensors with EPN which acts as a spectral detector of higher-order occurrences to prevent burstiness. We also prove that for a tensor of order r built from d dimensional feature descriptors, such a detector gives the likelihood if at least one higher-order occurrence is 'projected' into one of binom(d,r) subspaces represented by the tensor; thus forming a tensor power normalization metric endowed with binom(d,r) such 'detectors'. For experimental contributions, we apply several second- and higher-order pooling variants to action recognition, provide previously not presented comparisons of such pooling variants, and show state-of-the-art results on HMDB-51, YUP++ and MPII Cooking Activities., Comment: Accepted by IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2024)

The theoretical advantages of distributional reinforcement learning~(RL) over classical RL remain elusive despite its remarkable empirical performance. Starting from Categorical Distributional RL~(CDRL), we attribute the potential superiority of distributional RL to a derived distribution-matching regularization by applying a return density function decomposition technique. This unexplored regularization in the distributional RL context is aimed at capturing additional return distribution information regardless of only its expectation, contributing to an augmented reward signal in the policy optimization. Compared with the entropy regularization in MaxEnt RL that explicitly optimizes the policy to encourage the exploration, the resulting regularization in CDRL implicitly optimizes policies guided by the new reward signal to align with the uncertainty of target return distributions, leading to an uncertainty-aware exploration effect. Finally, extensive experiments substantiate the importance of this uncertainty-aware regularization in distributional RL on the empirical benefits over classical RL.

Abnormal states in deep reinforcement learning~(RL) are states that are beyond the scope of an RL policy. Such states may lead to sub-optimal and unsafe decision making for the RL system, impeding its deployment in real scenarios. In this paper, we propose a simple yet effective anomaly detection framework for deep RL algorithms that simultaneously considers random, adversarial and out-of-distribution~(OOD) state outliers. In particular, we attain the class-conditional distributions for each action class under the Gaussian assumption, and rely on these distributions to discriminate between inliers and outliers based on Mahalanobis Distance~(MD) and Robust Mahalanobis Distance. We conduct extensive experiments on Atari games that verify the effectiveness of our detection strategies. To the best of our knowledge, we present the first in-detail study of statistical and adversarial anomaly detection in deep RL algorithms. This simple unified anomaly detection paves the way towards deploying safe RL systems in real-world applications., Comment: 19 pages, 21 figures

In real scenarios, state observations that an agent observes may contain measurement errors or adversarial noises, misleading the agent to take suboptimal actions or even collapse while training. In this paper, we study the training robustness of distributional Reinforcement Learning (RL), a class of state-of-the-art methods that estimate the whole distribution, as opposed to only the expectation, of the total return. Firstly, we validate the contraction of distributional Bellman operators in the State-Noisy Markov Decision Process (SN-MDP), a typical tabular case that incorporates both random and adversarial state observation noises. In the noisy setting with function approximation, we then analyze the vulnerability of least squared loss in expectation-based RL with either linear or nonlinear function approximation. By contrast, we theoretically characterize the bounded gradient norm of distributional RL loss based on the categorical parameterization equipped with the KL divergence. The resulting stable gradients while the optimization in distributional RL accounts for its better training robustness against state observation noises. Finally, extensive experiments on the suite of environments verified that distributional RL is less vulnerable against both random and adversarial noisy state observations compared with its expectation-based counterpart., Comment: Accepted in ECML PKDD 2023. This is the authors version of the work. The definitive Version of Record will be published in the Proceedings of ECML PKDD 2023

In the realm of deep learning, the Fisher information matrix (FIM) gives novel insights and useful tools to characterize the loss landscape, perform second-order optimization, and build geometric learning theories. The exact FIM is either unavailable in closed form or too expensive to compute. In practice, it is almost always estimated based on empirical samples. We investigate two such estimators based on two equivalent representations of the FIM -- both unbiased and consistent. Their estimation quality is naturally gauged by their variance given in closed form. We analyze how the parametric structure of a deep neural network can affect the variance. The meaning of this variance measure and its upper bounds are then discussed in the context of deep learning., Comment: Published in Advances in Neural Information Processing Systems 34 (NeurIPS 2021)

We implement training of neural networks in secure multi-party computation (MPC) using quantization commonly used in said setting. We are the first to present an MNIST classifier purely trained in MPC that comes within 0.2 percent of the accuracy of the same convolutional neural network trained via plaintext computation. More concretely, we have trained a network with two convolutional and two dense layers to 99.2% accuracy in 3.5 hours (under one hour for 99% accuracy). We have also implemented AlexNet for CIFAR-10, which converges in a few hours. We develop novel protocols for exponentiation and inverse square root. Finally, we present experiments in a range of MPC security models for up to ten parties, both with honest and dishonest majority as well as semi-honest and malicious security., Comment: 27 pages

Graphs are widely used as a popular representation of the network structure of connected data. Graph data can be found in a broad spectrum of application domains such as social systems, ecosystems, biological networks, knowledge graphs, and information systems. With the continuous penetration of artificial intelligence technologies, graph learning (i.e., machine learning on graphs) is gaining attention from both researchers and practitioners. Graph learning proves effective for many tasks, such as classification, link prediction, and matching. Generally, graph learning methods extract relevant features of graphs by taking advantage of machine learning algorithms. In this survey, we present a comprehensive overview on the state-of-the-art of graph learning. Special attention is paid to four categories of existing graph learning methods, including graph signal processing, matrix factorization, random walk, and deep learning. Major models and algorithms under these categories are reviewed respectively. We examine graph learning applications in areas such as text, images, science, knowledge graphs, and combinatorial optimization. In addition, we discuss several promising research directions in this field., Comment: 19 pages, 6 figures

Features representation leverages the great power in network analysis tasks. However, most features are discrete which poses tremendous challenges to effective use. Recently, increasing attention has been paid on network feature learning, which could map discrete features to continued space. Unfortunately, current studies fail to fully preserve the structural information in the feature space due to random negative sampling strategy during training. To tackle this problem, we study the problem of feature learning and novelty propose a force-based graph learning model named GForce inspired by the spring-electrical model. GForce assumes that nodes are in attractive forces and repulsive forces, thus leading to the same representation with the original structural information in feature learning. Comprehensive experiments on benchmark datasets demonstrate the effectiveness of the proposed framework. Furthermore, GForce opens up opportunities to use physics models to model node interaction for graph learning.

Network representation learning (NRL) is an effective graph analytics technique and promotes users to deeply understand the hidden characteristics of graph data. It has been successfully applied in many real-world tasks related to network science, such as social network data processing, biological information processing, and recommender systems. Deep Learning is a powerful tool to learn data features. However, it is non-trivial to generalize deep learning to graph-structured data since it is different from the regular data such as pictures having spatial information and sounds having temporal information. Recently, researchers proposed many deep learning-based methods in the area of NRL. In this survey, we investigate classical NRL from traditional feature learning method to the deep learning-based model, analyze relationships between them, and summarize the latest progress. Finally, we discuss open issues considering NRL and point out the future directions in this field.

Softmax is widely used in deep learning to map some representation to a probability distribution. As it is based on exp/log functions that are relatively expensive in multi-party computation, Mohassel and Zhang (2017) proposed a simpler replacement based on ReLU to be used in secure computation. However, we could not reproduce the accuracy they reported for training on MNIST with three fully connected layers. Later works (e.g., Wagh et al., 2019 and 2021) used the softmax replacement not for computing the output probability distribution but for approximating the gradient in back-propagation. In this work, we analyze the two uses of the replacement and compare them to softmax, both in terms of accuracy and cost in multi-party computation. We found that the replacement only provides a significant speed-up for a one-layer network while it always reduces accuracy, sometimes significantly. Thus we conclude that its usefulness is limited and one should use the original softmax function instead., Comment: 6 pages, PPML/PriML workshop at NeurIPS 2020; updated accuracy figures after bug fix

Multivariate relations are general in various types of networks, such as biological networks, social networks, transportation networks, and academic networks. Due to the principle of ternary closures and the trend of group formation, the multivariate relationships in social networks are complex and rich. Therefore, in graph learning tasks of social networks, the identification and utilization of multivariate relationship information are more important. Existing graph learning methods are based on the neighborhood information diffusion mechanism, which often leads to partial omission or even lack of multivariate relationship information, and ultimately affects the accuracy and execution efficiency of the task. To address these challenges, this paper proposes the multivariate relationship aggregation learning (MORE) method, which can effectively capture the multivariate relationship information in the network environment. By aggregating node attribute features and structural features, MORE achieves higher accuracy and faster convergence speed. We conducted experiments on one citation network and five social networks. The experimental results show that the MORE model has higher accuracy than the GCN (Graph Convolutional Network) model in node classification tasks, and can significantly reduce time cost., Comment: 11 pages, 6 figures

The scarcity of class-labeled data is a ubiquitous bottleneck in many machine learning problems. While abundant unlabeled data typically exist and provide a potential solution, it is highly challenging to exploit them. In this paper, we address this problem by leveraging Positive-Unlabeled~(PU) classification and the conditional generation with extra unlabeled data \emph{simultaneously}. In particular, we present a novel training framework to jointly target both PU classification and conditional generation when exposed to extra data, especially out-of-distribution unlabeled data, by exploring the interplay between them: 1) enhancing the performance of PU classifiers with the assistance of a novel Classifier-Noise-Invariant Conditional GAN~(CNI-CGAN) that is robust to noisy labels, 2) leveraging extra data with predicted labels from a PU classifier to help the generation. Theoretically, we prove the optimal condition of CNI-CGAN, and experimentally, we conducted extensive evaluations on diverse datasets, verifying the simultaneous improvements in both classification and generation.

This letter introduces an abstract learning problem called the "set embedding": The objective is to map sets into probability distributions so as to lose less information. We relate set union and intersection operations with corresponding interpolations of probability distributions. We also demonstrate a preliminary solution with experimental results on toy set embedding examples., Comment: To be presented at Sets & Partitions (NeurIPS 2019 workshop)

Regularization plays a crucial role in machine learning models, especially for deep neural networks. The existing regularization techniques mainly rely on the i.i.d. assumption and only consider the knowledge from the current sample, without the leverage of the neighboring relationship between samples. In this work, we propose a general regularizer called \textbf{Patch-level Neighborhood Interpolation~(Pani)} that conducts a non-local representation in the computation of networks. Our proposal explicitly constructs patch-level graphs in different layers and then linearly interpolates neighborhood patch features, serving as a general and effective regularization strategy. Further, we customize our approach into two kinds of popular regularization methods, namely Virtual Adversarial Training (VAT) and MixUp as well as its variants. The first derived \textbf{Pani VAT} presents a novel way to construct non-local adversarial smoothness by employing patch-level interpolated perturbations. The second derived \textbf{Pani MixUp} method extends the MixUp, and achieves superiority over MixUp and competitive performance over state-of-the-art variants of MixUp method with a significant advantage in computational efficiency. Extensive experiments have verified the effectiveness of our Pani approach in both supervised and semi-supervised settings., Comment: Accepted in ACML 2023 conference track

iDASH is a competition soliciting implementations of cryptographic schemes of interest in the context of biology. In 2019, one track asked for multi-party computation implementations of training of a machine learning model suitable for two datasets from cancer research. In this note, we describe our solution submitted to the competition. We found that the training can be run on three AWS c5.9xlarge instances in less then one minute using MPC tolerating one semi-honest corruption, and less than ten seconds at a slightly lower accuracy., Comment: 4 pages

The design of deep graph models still remains to be investigated and the crucial part is how to explore and exploit the knowledge from different hops of neighbors in an efficient way. In this paper, we propose a novel RNN-like deep graph neural network architecture by incorporating AdaBoost into the computation of network; and the proposed graph convolutional network called AdaGCN~(Adaboosting Graph Convolutional Network) has the ability to efficiently extract knowledge from high-order neighbors of current nodes and then integrates knowledge from different hops of neighbors into the network in an Adaboost way. Different from other graph neural networks that directly stack many graph convolution layers, AdaGCN shares the same base neural network architecture among all ``layers'' and is recursively optimized, which is similar to an RNN. Besides, We also theoretically established the connection between AdaGCN and existing graph convolutional methods, presenting the benefits of our proposal. Finally, extensive experiments demonstrate the consistent state-of-the-art prediction performance on graphs across different label rates and the computational advantage of our approach AdaGCN~\footnote{Code is available at \url{https://github.com/datake/AdaGCN}.}, Comment: Published on International Conference on Learning Representations (ICLR) 2021

Why do deep neural networks (DNNs) benefit from very high dimensional parameter spaces? Their huge parameter complexities vs. stunning performances in practice is all the more intriguing and not explainable using the standard theory of model selection for regular models. In this work, we propose a geometrically flavored information-theoretic approach to study this phenomenon. Namely, we introduce the locally varying dimensionality of the parameter space of neural network models by considering the number of significant dimensions of the Fisher information matrix, and model the parameter space as a manifold using the framework of singular semi-Riemannian geometry. We derive model complexity measures which yield short description lengths for deep neural network models based on their singularity analysis thus explaining the good performance of DNNs despite their large number of parameters., Comment: This work first appeared under the former title "Lightlike Neuromanifolds, Occam's Razor and Deep Learning"

In a graph convolutional network, we assume that the graph $G$ is generated wrt some observation noise. During learning, we make small random perturbations $\Delta{}G$ of the graph and try to improve generalization. Based on quantum information geometry, $\Delta{}G$ can be characterized by the eigendecomposition of the graph Laplacian matrix. We try to minimize the loss wrt the perturbed $G+\Delta{G}$ while making $\Delta{G}$ to be effective in terms of the Fisher information of the neural network. Our proposed model can consistently improve graph convolutional networks on semi-supervised node classification tasks with reasonable computational overhead. We present three different geometries on the manifold of graphs: the intrinsic geometry measures the information theoretic dynamics of a graph; the extrinsic geometry characterizes how such dynamics can affect externally a graph neural network; the embedding geometry is for measuring node embeddings. These new analytical tools are useful in developing a good understanding of graph neural networks and fostering new techniques., Comment: Published in UAI 2019

Deep neural networks have been widely deployed in various machine learning tasks. However, recent works have demonstrated that they are vulnerable to adversarial examples: carefully crafted small perturbations to cause misclassification by the network. In this work, we propose a novel defense mechanism called Boundary Conditional GAN to enhance the robustness of deep neural networks against adversarial examples. Boundary Conditional GAN, a modified version of Conditional GAN, can generate boundary samples with true labels near the decision boundary of a pre-trained classifier. These boundary samples are fed to the pre-trained classifier as data augmentation to make the decision boundary more robust. We empirically show that the model improved by our approach consistently defenses against various types of adversarial attacks successfully. Further quantitative investigations about the improvement of robustness and visualization of decision boundaries are also provided to justify the effectiveness of our strategy. This new defense mechanism that uses boundary samples to enhance the robustness of networks opens up a new way to defense adversarial attacks consistently.

Most previous works usually explained adversarial examples from several specific perspectives, lacking relatively integral comprehension about this problem. In this paper, we present a systematic study on adversarial examples from three aspects: the amount of training data, task-dependent and model-specific factors. Particularly, we show that adversarial generalization (i.e. test accuracy on adversarial examples) for standard training requires more data than standard generalization (i.e. test accuracy on clean examples); and uncover the global relationship between generalization and robustness with respect to the data size especially when data is augmented by generative models. This reveals the trade-off correlation between standard generalization and robustness in limited training data regime and their consistency when data size is large enough. Furthermore, we explore how different task-dependent and model-specific factors influence the vulnerability of deep neural networks by extensive empirical analysis. Relevant recommendations on defense against adversarial attacks are provided as well. Our results outline a potential path towards the luminous and systematic understanding of adversarial examples.

Graph Convolutional Networks(GCNs) play a crucial role in graph learning tasks, however, learning graph embedding with few supervised signals is still a difficult problem. In this paper, we propose a novel training algorithm for Graph Convolutional Network, called Multi-Stage Self-Supervised(M3S) Training Algorithm, combined with self-supervised learning approach, focusing on improving the generalization performance of GCNs on graphs with few labeled nodes. Firstly, a Multi-Stage Training Framework is provided as the basis of M3S training method. Then we leverage DeepCluster technique, a popular form of self-supervised learning, and design corresponding aligning mechanism on the embedding space to refine the Multi-Stage Training Framework, resulting in M3S Training Algorithm. Finally, extensive experimental results verify the superior performance of our algorithm on graphs with few labeled nodes under different label rates compared with other state-of-the-art approaches., Comment: AAAI Conference on Artificial Intelligence (AAAI 2020)

The effectiveness of Graph Convolutional Networks (GCNs) has been demonstrated in a wide range of graph-based machine learning tasks. However, the update of parameters in GCNs is only from labeled nodes, lacking the utilization of unlabeled data. In this paper, we apply Virtual Adversarial Training (VAT), an adversarial regularization method based on both labeled and unlabeled data, on the supervised loss of GCN to enhance its generalization performance. By imposing virtually adversarial smoothness on the posterior distribution in semi-supervised learning, VAT yields improvement on the Symmetrical Laplacian Smoothness of GCNs. In addition, due to the difference of property in features, we perturb virtual adversarial perturbations on sparse and dense features, resulting in GCN Sparse VAT (GCNSVAT) and GCN Dense VAT (GCNDVAT) algorithms, respectively. Extensive experiments verify the effectiveness of our two methods across different training sizes. Our work paves the way towards better understanding the direction of improvement on GCNs in the future., Comment: Chinese Conference on Pattern Recognition and Computer Vision(PRCV) 2019 Oral paper

We define a novel class of distances between statistical multivariate distributions by modeling an optimal transport problem on their marginals with respect to a ground distance defined on their conditionals. These new distances are metrics whenever the ground distance between the marginals is a metric, generalize both the Wasserstein distances between discrete measures and a recently introduced metric distance between statistical mixtures, and provide an upper bound for jointly convex distances between statistical mixtures. By entropic regularization of the optimal transport, we obtain a fast differentiable Sinkhorn-type distance. We experimentally evaluate our new family of distances by quantifying the upper bounds of several jointly convex distances between statistical mixtures, and by proposing a novel efficient method to learn Gaussian mixture models (GMMs) by simplifying kernel density estimators with respect to our distance. Our GMM learning technique experimentally improves significantly over the EM implementation of {\tt sklearn} on the {\tt MNIST} and {\tt Fashion MNIST} datasets., Comment: 23 pages, 6 figures

A basic problem in machine learning is to find a mapping $f$ from a low dimensional latent space $\mathcal{Y}$ to a high dimensional observation space $\mathcal{X}$. Modern tools such as deep neural networks are capable to represent general non-linear mappings. A learner can easily find a mapping which perfectly fits all the observations. However, such a mapping is often not considered as good, because it is not simple enough and can overfit. How to define simplicity? We try to make a formal definition on the amount of information imposed by a non-linear mapping $f$. Intuitively, we measure the local discrepancy between the pullback geometry and the intrinsic geometry of the latent space. Our definition is based on information geometry and is independent of the empirical observations, nor specific parameterizations. We prove its basic properties and discuss relationships with related machine learning methods.

The total variation distance is a core statistical distance between probability measures that satisfies the metric axioms, with value always falling in $[0,1]$. This distance plays a fundamental role in machine learning and signal processing: It is a member of the broader class of $f$-divergences, and it is related to the probability of error in Bayesian hypothesis testing. Since the total variation distance does not admit closed-form expressions for statistical mixtures (like Gaussian mixture models), one often has to rely in practice on costly numerical integrations or on fast Monte Carlo approximations that however do not guarantee deterministic lower and upper bounds. In this work, we consider two methods for bounding the total variation of univariate mixture models: The first method is based on the information monotonicity property of the total variation to design guaranteed nested deterministic lower bounds. The second method relies on computing the geometric lower and upper envelopes of weighted mixture components to derive deterministic bounds based on density ratio. We demonstrate the tightness of our bounds in a series of experiments on Gaussian, Gamma and Rayleigh mixture models., Comment: 11 pages, 2 figures

We propose a new generic type of stochastic neurons, called $q$-neurons, that considers activation functions based on Jackson's $q$-derivatives with stochastic parameters $q$. Our generalization of neural network architectures with $q$-neurons is shown to be both scalable and very easy to implement. We demonstrate experimentally consistently improved performances over state-of-the-art standard activation functions, both on training and testing loss functions., Comment: 12 pages, 5 figures, 1 table

Clustering categorical distributions in the finite-dimensional probability simplex is a fundamental task met in many applications dealing with normalized histograms. Traditionally, the differential-geometric structures of the probability simplex have been used either by (i) setting the Riemannian metric tensor to the Fisher information matrix of the categorical distributions, or (ii) defining the dualistic information-geometric structure induced by a smooth dissimilarity measure, the Kullback-Leibler divergence. In this work, we introduce for clustering tasks a novel computationally-friendly framework for modeling geometrically the probability simplex: The {\em Hilbert simplex geometry}. In the Hilbert simplex geometry, the distance is the non-separable Hilbert's metric distance which satisfies the property of information monotonicity with distance level set functions described by polytope boundaries. We show that both the Aitchison and Hilbert simplex distances are norm distances on normalized logarithmic representations with respect to the $\ell_2$ and variation norms, respectively. We discuss the pros and cons of those different statistical modelings, and benchmark experimentally these different kind of geometries for center-based $k$-means and $k$-center clustering. Furthermore, since a canonical Hilbert distance can be defined on any bounded convex subset of the Euclidean space, we also consider Hilbert's geometry of the elliptope of correlation matrices and study its clustering performances compared to Fr\"obenius and log-det divergences., Comment: 48 pages