Beyond its highly publicized victories in Go, there have been numerous successful applications of deep learning in information retrieval, computer vision and speech recognition. In cybersecurity, an increasing number of companies have become excited about the potential of deep learning, and have started to use it for various security incidents, the most popular being malware detection.ese companies assert that deep learning (DL) could help turn the tide in the ba le against malware infections. However, deep neural networks (DNNs) are vulnerable to adversarial samples, a aw that plagues most if not all statistical learning models. Recent research has demonstrated that those with malicious intent can easily circumvent deep learning-powered malware detection by exploiting this aw.In order to address this problem, previous work has developed various defense mechanisms that either augmenting training data or enhance model's complexity. However, a er a thorough analysis of the fundamental aw in DNNs, we discover that the e ectiveness of current defenses is limited and, more importantly, cannot provide theoretical guarantees as to their robustness against adversarial sampled-based a acks. As such, we propose a new adversary resistant technique that obstructs a ackers from constructing impactful adversarial samples by randomly nullifying features within samples. In this work, we evaluate our proposed technique against a real world dataset with 14,679 malware variants and 17,399 benign programs. We theoretically validate the robustness of our technique, and empirically show that our technique signi cantly boosts DNN robustness to adversarial samples while maintaining high accuracy in classi cation. To demonstrate the general applicability of our proposed method, we also conduct experiments using the MNIST and CIFAR-10 datasets, generally used in image recognition research. ACM Reference format:

We propose novel neural temporal models for predicting and synthesizing human motion, achieving state-of-theart in modeling long-term motion trajectories while being competitive with prior work in short-term prediction, with significantly less required computation. Key aspects of our proposed system include: 1) a novel, two-level processing architecture that aids in generating planned trajectories, 2) a simple set of easily computable features that integrate derivative information into the model, and 3) a novel multi-objective loss function that helps the model to slowly progress from the simpler task of next-step prediction to the harder task of multi-step closed-loop prediction. Our results demonstrate that these innovations facilitate improved modeling of long-term motion trajectories. Finally, we propose a novel metric, called Normalized Power Spectrum Similarity (NPSS), to evaluate the long-term predictive ability of motion synthesis models, complementing the popular mean-squared error (MSE) measure of the Euler joint angles over time. We conduct a user study to determine if the proposed NPSS correlates with human evaluation of longterm motion more strongly than MSE and find that it indeed does.

Rule extraction from black box models is critical in domains that require model validation before implementation, as can be the case in credit scoring and medical diagnosis. Though already a challenging problem in statistical learning in general, the difficulty is even greater when highly nonlinear, recursive models, such as recurrent neural networks (RNNs), are fit to data. Here, we study the extraction of rules from second-order RNNs trained to recognize the Tomita grammars. We show that production rules can be stably extracted from trained RNNs and that in certain cases, the rules outperform the trained RNNs.

Finding biologically plausible alternatives to back-propagation of errors is a fundamentally important challenge in artificial neural network research. In this paper, we propose a learning algorithm called error-driven Local Representation Alignment (LRA-E), which has strong connections to predictive coding, a theory that offers a mechanistic way of describing neurocomputational machinery. In addition, we propose an improved variant of Difference Target Propagation, another procedure that comes from the same family of algorithms as LRA-E. We compare our procedures to several other biologicallymotivated algorithms, including two feedback alignment algorithms and Equilibrium Propagation. In two benchmarks, we find that both of our proposed algorithms yield stable performance and strong generalization compared to other competing back-propagation alternatives when training deeper, highly nonlinear networks, with LRA-E performing the best overall.

No abstract