Author Description

Machine learning has become one of the most exciting areas in recent years, as it has achieved state-of-the-art performance and demonstrated fundamental breakthroughs in many challenging tasks. The efficiency and scalability of a machine learning system are important and restrict the applicability of the system. We study the problem of learning from group comparisons, with applications in predicting outcomes of sports and online games. In this work, we propose a new model that takes the player-interaction effects into consideration and show that our proposed models have much better prediction power on several E-sports datasets, and furthermore can be used to reveal interesting patterns that cannot be discovered by previous methods. Despite the efficiency and scalability of machine learning systems, recent studies have demonstrated that many machine learning models, especially deep neural networks, are vulnerable to adversarial examples, i.e., examples that are carefully crafted to fool a well-trained machine learning model while being indistinguishable from the natural images to humans. This makes it unsafe to apply deep neural networks in security-critical areas. We present two algorithms, Embedding Regularized Classifier (ER-Classifier) and Adversarial Example Detector (ADEtector), to improve the robustness of deep neural networks against adversarial examples. Motivated by the observation that adversarial examples often lie outside the natural image data manifold and the intrinsic dimension of image data is much smaller than its pixel space dimension, we propose to embed high-dimensional input images into a low-dimensional space and apply regularization on the embedding space to push the adversarial examples back to the manifold. Another algorithm, ADEtector, is inspired by the similar ideas of ER-Classifier and the fact that detecting adversarial examples are easier than classifying them correctly, we propose a novel adversarial example detector, short for ADEtec