Adversarial examples have become one of the largest challenges that machine learning models, especially neural network classifiers, face. These adversarial examples break the assumption of attack-free scenario and fool state-of-the-art (SOTA) classifiers with insignificant perturbations to human. So far, researchers achieved great progress in utilizing adversarial training as a defense. However, the overwhelming computational cost degrades its applicability and little has been done to overcome this issue. Single-Step adversarial training methods have been proposed as computationally viable solutions, however they still fail to defend against iterative adversarial examples. In this work, we first experimentally analyze several different SOTA defense methods against adversarial examples. Then, based on observations from experiments, we propose a novel single-step adversarial training method which can defend against both single-step and iterative adversarial examples. Lastly, through extensive evaluations, we demonstrate that our proposed method outperforms the SOTA single-step and iterative adversarial training defense. Compared with ATDA (single-step method) on CIFAR10 dataset, our proposed method achieves 35.67% enhancement in test accuracy and 19.14% reduction in training time. When compared with methods that use BIM or Madry examples (iterative methods) on CIFAR10 dataset, it saves up to 76.03% in training time with less than 3.78% degeneration in test accuracy.

Higher-order Recurrent Neural Networks (RNNs) are effective for long-term forecasting since such architectures can model higher-order correlations and long-term dynamics more effectively. However, higher-order models are expensive and require exponentially more parameters and operations compared with their first-order counterparts. This problem is particularly pronounced in multidimensional data such as videos. To address this issue, we propose Convolutional Tensor-Train Decomposition (CTTD), a novel tensor decomposition with convolutional operations. With CTTD, we construct Convolutional Tensor-Train LSTM (Conv-TT-LSTM) to capture higher-order space-time correlations in videos. We demonstrate that the proposed model outperforms the conventional (first-order) Convolutional LSTM (ConvLSTM) as well as the state-of-the-art ConvLSTM-based approaches in pixel-level video prediction tasks on Moving-MNIST and KTH action datasets, but with much fewer parameters.

Image classification requires the generation of features capable of detecting image patterns informative of group identity. The objective of this study was to classify images from the public CIFAR-10 image dataset by leveraging combinations of disparate image feature sources from both manual and deep learning approaches. Histogram of oriented gradients (HOG) and pixel intensities successfully inform classification (53% and 59% classification accuracy, respectively), yet there is much room for improvement. VGG16 with ImageNet trained weights and a CIFAR-10 optimized model (CIFAR-VGG) further improve upon image classification (60% and 93.43% accuracy, respectively). We further improved classification by utilizing transfer learning to re-establish optimal network weights for VGG16 (TL-VGG) and Inception ResNet v2 (TL-Inception) resulting in significant performance increases (85% and 90.74%, respectively), yet fail to surpass CIFAR-VGG. We hypothesized that if each generated feature set obtained some unique insight into the classification problem, then combining these features would result in greater classification accuracy, surpassing that of CIFAR-VGG. Upon selection of the top 1000 principal components from TL-VGG, TL-Inception, HOG, pixel intensities, and CIFAR-VGG, we achieved testing accuracy of 94.6%, lending support to our hypothesis.

A new algorithmic framework is proposed for learning autoencoders of data distributions. We minimize the discrepancy between the model and target distributions, with a \emph{relational regularization} on the learnable latent prior. This regularization penalizes the fused Gromov-Wasserstein (FGW) distance between the latent prior and its corresponding posterior, allowing one to flexibly learn a structured prior distribution associated with the generative model. Moreover, it helps co-training of multiple autoencoders even if they have heterogeneous architectures and incomparable latent spaces. We implement the framework with two scalable algorithms, making it applicable for both probabilistic and deterministic autoencoders. Our relational regularized autoencoder (RAE) outperforms existing methods, $e.g.$, the variational autoencoder, Wasserstein autoencoder, and their variants, on generating images. Additionally, our relational co-training strategy for autoencoders achieves encouraging results in both synthesis and real-world multi-view learning tasks.

One-shot neural architecture search allows joint learning of weights and network architecture, reducing computational cost. We limit our search space to the depth of residual networks and formulate an analytically tractable variational objective that allows for obtaining an unbiased approximate posterior over depths in one-shot. We propose a heuristic to prune our networks based on this distribution. We compare our proposed method against manual search over network depths on the MNIST, Fashion-MNIST, SVHN datasets. We find that pruned networks do not incur a loss in predictive performance, obtaining accuracies competitive with unpruned networks. Marginalising over depth allows us to obtain better-calibrated test-time uncertainty estimates than regular networks, in a single forward pass.

Community structure in networks is observed in many different domains, and unsupervised community detection has received a lot of attention in the literature. Increasingly the focus of network analysis is shifting towards using network information in some other prediction or inference task rather than just analyzing the network itself. In particular, in neuroimaging applications brain networks are available for multiple subjects and the goal is often to predict a phenotype of interest. Community structure is well known to be a feature of brain networks, typically corresponding to different regions of the brain responsible for different functions. There are standard parcellations of the brain into such regions, usually obtained by applying clustering methods to brain connectomes of healthy subjects. However, when the goal is predicting a phenotype or distinguishing between different conditions, these static communities from an unrelated set of healthy subjects may not be the most useful for prediction. Here we present a method for supervised community detection, aiming to find a partition of the network into communities that is most useful for predicting a particular response. We use a block-structured regularization penalty combined with a prediction loss function, and compute the solution with a combination of a spectral method and an ADMM optimization algorithm. We show that the spectral clustering method recovers the correct communities under a weighted stochastic block model. The method performs well on both simulated and real brain networks, providing support for the idea of task-dependent brain regions.

Human gaze is known to be an intention-revealing signal in human demonstrations of tasks. In this work, we use gaze cues from human demonstrators to enhance the performance of state-of-the-art inverse reinforcement learning (IRL) and behavior cloning (BC) algorithms. We propose a novel approach for utilizing gaze data in a computationally efficient manner --- encoding the human's attention as part of an auxiliary loss function, without adding any additional learnable parameters to those models and without requiring gaze data at test time. The auxiliary loss encourages a network to have convolutional activations in regions where the human's gaze fixated. We show how to augment any existing convolutional architecture with our auxiliary gaze loss (coverage-based gaze loss or CGL) that can guide learning toward a better reward function or policy. We show that our proposed approach consistently improves performance of both BC and IRL methods on a variety of Atari games. We also compare against two baseline methods for utilizing gaze data with imitation learning methods. Our approach outperforms a baseline method, called gaze-modulated dropout (GMD), and is comparable to another method (AGIL) which uses gaze as input to the network and thus increases the amount of learnable parameters.

Model-free deep reinforcement learning algorithms are troubled with poor sample efficiency -- learning reliable policies generally requires a vast amount of interaction with the environment. One hypothesis is that catastrophic interference between various segments within the environment is an issue. In this paper, we perform a large-scale empirical study on the presence of catastrophic interference in the Arcade Learning Environment and find that learning particular game segments frequently degrades performance on previously learned segments. In what we term the Memento observation, we show that an identically parameterized agent spawned from a state where the original agent plateaued, reliably makes further progress. This phenomenon is general -- we find consistent performance boosts across architectures, learning algorithms and environments. Our results indicate that eliminating catastrophic interference can contribute towards improved performance and data efficiency of deep reinforcement learning algorithms.

We study the estimation of risk-sensitive policies in reinforcement learning problems defined by a Markov Decision Process (MDPs) whose state and action spaces are countably finite. Prior efforts are predominately afflicted by computational challenges associated with the fact that risk-sensitive MDPs are time-inconsistent. To ameliorate this issue, we propose a new definition of risk, which we call caution, as a penalty function added to the dual objective of the linear programming (LP) formulation of reinforcement learning. The caution measures the distributional risk of a policy, which is a function of the policy's long-term state occupancy distribution. To solve this problem in an online model-free manner, we propose a stochastic variant of primal-dual method that uses Kullback-Lieber (KL) divergence as its proximal term. We establish that the number of iterations/samples required to attain approximately optimal solutions of this scheme matches tight dependencies on the cardinality of the state and action spaces, but differs in its dependence on the infinity norm of the gradient of the risk measure. Experiments demonstrate the merits of this approach for improving the reliability of reward accumulation without additional computational burdens.

This paper presents a data structure that summarizes distances between configurations across a robot configuration space, using a binary space partition whose cells contain parameters used for a locally linear approximation of the distance function. Querying the data structure is extremely fast, particularly when compared to the graph search required for querying Probabilistic Roadmaps, and memory requirements are promising. The paper explores the use of the data structure constructed for a single robot to provide a heuristic for challenging multi-robot motion planning problems. Potential applications also include the use of remote computation to analyze the space of robot motions, which then might be transmitted on-demand to robots with fewer computational resources.