(Frankle & Carbin, 2019) shows that there exist winning tickets (small but critical subnetworks) for dense, randomly initialized networks, that can be trained alone to achieve comparable accuracies to the latter in a similar number of iterations. However, the identification of these winning tickets still requires the costly train-prune-retrain process, limiting their practical benefits. In this paper, we discover for the first time that the winning tickets can be identified at the very early training stage, which we term as early-bird (EB) tickets, via low-cost training schemes (e.g., early stopping and low-precision training) at large learning rates. Our finding of EB tickets is consistent with recently reported observations that the key connectivity patterns of neural networks emerge early. Furthermore, we propose a mask distance metric that can be used to identify EB tickets with low computational overhead, without needing to know the true winning tickets that emerge after the full training. Finally, we leverage the existence of EB tickets and the proposed mask distance to develop efficient training methods, which are achieved by first identifying EB tickets via low-cost schemes, and then continuing to train merely the EB tickets towards the target accuracy. Experiments based on various deep networks and datasets validate: 1) the existence of EB tickets, and the effectiveness of mask distance in efficiently identifying them; and 2) that the proposed efficient training via EB tickets can achieve up to 4.7x energy savings while maintaining comparable or even better accuracy, demonstrating a promising and easily adopted method for tackling cost-prohibitive deep network training.

In hyperspectral image (HSI) classification, spatial context has demonstrated its significance in achieving promising performance. However, conventional spatial context-based methods simply assume that spatially neighboring pixels should correspond to the same land-cover class, so they often fail to correctly discover the contextual relations among pixels in complex situations, and thus leading to imperfect classification results on some irregular or inhomogeneous regions such as class boundaries. To address this deficiency, we develop a new HSI classification method based on the recently proposed Graph Convolutional Network (GCN), as it can flexibly encode the relations among arbitrarily structured non-Euclidean data. Different from traditional GCN, there are two novel strategies adopted by our method to further exploit the contextual relations for accurate HSI classification. First, since the receptive field of traditional GCN is often limited to fairly small neighborhood, we proposed to capture long range contextual relations in HSI by performing successive graph convolutions on a learned region-induced graph which is transformed from the original 2D image grids. Second, we refine the graph edge weight and the connective relationships among image regions by learning the improved adjacency matrix and the 'edge filter', so that the graph can be gradually refined to adapt to the representations generated by each graph convolutional layer. Such updated graph will in turn result in accurate region representations, and vice versa. The experiments carried out on three real-world benchmark datasets demonstrate that the proposed method yields significant improvement in the classification performance when compared with some state-of-the-art approaches.

A long-standing challenge in reinforcement learning is the design of function approximations and efficient learning algorithms that provide agents with fast training, robust learning , and high performance in complex environments. To this end, the use of prior knowledge, while promising, is often costly and, in essence, challenging to scale up. In contrast, we consider problem knowledge signals, that are any relevant indicator useful to solve a task, e.g., metrics of uncertainty or proactive prediction of future states. Our framework consists of predicting such complementary quantities associated with self-performance assessment and accurate expectations. Therefore, policy and value functions are no longer only optimized for a reward but are learned using environment-agnostic quantities. We propose a generally applicable framework for structuring reinforcement learning by injecting problem knowledge in policy gradient updates. In this paper: (a) We introduce MERL, our multi-head reinforcement learning framework for generalized auxiliary tasks. (b) We conduct experiments across a variety of standard benchmark environments. Our results show that MERL improves performance for on-and off-policy methods. (c) We show that MERL also improves transfer learning on a set of challenging tasks. (d) We investigate how our approach addresses the problem of reward sparsity and pushes the function approximations into a better-constrained parameter configuration.

Earlier methods for Neural Architecture Search were computationally expensive. Recently proposed Differentiable Neural Architecture Search algorithms such as DARTS can effectively speed up the computation. However, the current formulation relies on a relaxation of the original problem that leads to unstable and suboptimal solutions. We argue that these problems are caused by three fundamental reasons: (1) The difficulty of bi-level optimization; (2) Multicollinearity of correlated operations such as max pooling and average pooling; (3) The discrepancy between the optimization complexity of the search stage and the final training. In this paper, we propose a grouped variable pruning algorithm based on one-level optimization, which leads to a more stable and consistent optimization solution for differentiable NAS. Extensive experiments verify the superiority of the proposed method regarding both accuracy and stability. Our new approach obtains state-of-the-art accuracy on CIFAR-10, CIFAR-100 and ImageNet.

Gradient-based temporal difference (GTD) algorithms are widely used in off-policy learning scenarios. Among them, the two time-scale TD with gradient correction (TDC) algorithm has been shown to have superior performance. In contrast to previous studies that characterized the non-asymptotic convergence rate of TDC only under identical and independently distributed (i.i.d.) data samples, we provide the first non-asymptotic convergence analysis for two time-scale TDC under a non-i.i.d.\ Markovian sample path and linear function approximation. We show that the two time-scale TDC can converge as fast as O(log t/(t^(2/3))) under diminishing stepsize, and can converge exponentially fast under constant stepsize, but at the cost of a non-vanishing error. We further propose a TDC algorithm with blockwisely diminishing stepsize, and show that it asymptotically converges with an arbitrarily small error at a blockwisely linear convergence rate. Our experiments demonstrate that such an algorithm converges as fast as TDC under constant stepsize, and still enjoys comparable accuracy as TDC under diminishing stepsize.

ABSTRACT V oice activity detection (V AD), which classifies frames as speech or non-speech, is an important module in many speech applications including speaker verification. In this paper, we propose a novel method, called self-adaptive soft V AD, to incorporate a deep neural network (DNN)-based V AD into a deep speaker embedding system. The proposed method is a combination of the following two approaches. The first approach is soft V AD, which performs a soft selection of frame-level features extracted from a speaker feature extractor. The frame-level features are weighted by their corresponding speech posteriors estimated from the DNN-based V AD, and then aggregated to generate a speaker embedding. The second approach is self-adaptive V AD, which fine-tunes the pre-trained V AD on the speaker verification data to reduce the domain mismatch. Here, we introduce two unsupervised domain adaptation (DA) schemes, namely speech posterior-based DA (SP-DA) and joint learning-based DA (JL-DA). Experiments on a Korean speech database demonstrate that the verification performance is improved significantly in real-world environments by using self-adaptive soft V AD. Index T erms-- speaker verification, voice activity detection, unsupervised domain adaptation, soft V AD 1. INTRODUCTION Speaker verification (SV) is the task of verifying a person's claimed identity based on his or her voice.

Program verification offers a framework for ensuring program correctness and therefore systematically eliminating different classes of bugs. Inferring loop invariants is one of the main challenges behind automated verification of real-world programs which often contain many loops. In this paper, we present Continuous Logic Network (CLN), a novel neural architecture for automatically learning loop invariants directly from program execution traces. Unlike existing neural networks, CLNs can learn precise and explicit representations of formulas in Satisfiability Modulo Theories (SMT) for loop invariants from program execution traces. We develop a new sound and complete semantic mapping for assigning SMT formulas to continuous truth values that allows CLNs to be trained efficiently. We use CLNs to implement a new inference system for loop invariants, CLN2INV, that significantly outperforms existing approaches on the popular Code2Inv dataset. CLN2INV is the first tool to solve all 124 theoretically solvable problems in the Code2Inv dataset. Moreover, CLN2INV takes only 1.1 seconds on average for each problem, which is 40 times faster than existing approaches. We further demonstrate that CLN2INV can even learn 12 significantly more complex loop invariants than the ones required for the Code2Inv dataset.

One of the biggest issues in deep learning theory is the gener alization ability of networks with huge model size. The classical learning the ory suggests that overparameterized models cause overfitting. However, prac tically used large deep models avoid overfitting, which is not well explained by the c lassical approaches. To resolve this issue, several attempts have been made. Amon g them, the compression based bound is one of the promising approaches. However, the compression based bound can be applied only to a compressed network, and i t is not applicable to the non-compressed original network. In this paper, we gi ve a unified framework that can convert compression based bounds to those for n on-compressed original networks. The bound gives even better rate than the one for the compressed network by improving the bias term. By establishing the unified framework, we can obtain a data dependent generalization error bo und which gives a tighter evaluation than the data independent ones.

Abstract--Malware imposes tremendous threats to computer users nowadays. Since signature-based malware detection methods are neither effective nor efficient to identify new malware, many machine learning-based methods have been proposed. A common disadvantage of existing machine learning methods is that they are not based on understanding the full semantic meaning of assembly code of an executable. They rather use short assembly code fragments, because assembly code is usually too long to be modelled in its entirety . Another disadvantage is that those methods have either inferior performance or bad interpretability . T o overcome these challenges, we propose an Interpretable MAware Detector ( I-MAD), which achieves state-of-the-art performance on static malware detection with excellent interpretability . It integrates a hierarchical T ransformer network that can understand assembly code at the basic block, function, and executable level. It also integrates our novel interpretable feed-forward neural network to provide interpretations for its detection results by pointing out the impact of each feature with respect to the prediction. Experiment results show that our model significantly outperforms previous state-of-the-art static malware detection models and presents meaningful interpretations. Since the Internet has become an integral part of people's life, the large volume of malware spreading on it imposes tremendous threats to billions of netizens. Recognizing mal-ware samples downloaded by legitimate users in a timely manner is thus of crucial importance for their protection. Signature-based malware detection methods are widely used in antivirus products [1]. With the signatures extracted by malware analysts, known malware samples or some of their variants can be precisely recognized. However, with obfuscation techniques or even a change of compiler, it is easy to create variants of known malware that perform the same attack, but with literally different executable code. As a result, the previously crafted signatures can no longer recognize them [2]. Furthermore, signature-based detection is also ineffective to detect new and unseen malware in most cases.

We propose RecSim, a configurable platform for authoring simulation environments for recommender systems (RSs) that naturally supports sequential interaction with users. RecSim allows the creation of new environments that reflect particular aspects of user behavior and item structure at a level of abstraction well-suited to pushing the limits of current reinforcement learning (RL) and RS techniques in sequential interactive recommendation problems. Environments can be easily configured that vary assumptions about: user preferences and item familiarity; user latent state and its dynamics; and choice models and other user response behavior. We outline how RecSim offers value to RL and RS researchers and practitioners, and how it can serve as a vehicle for academic-industrial collaboration.