Deep Neural Networks (DNNs) have demonstrated remarkable performance in a diverse range of applications. Along with the prevalence of deep learning, it has been revealed that DNNs are vulnerable to attacks. By deliberately crafting adversarial examples, an adversary can manipulate a DNN to generate incorrect outputs, which may lead catastrophic consequences in applications such as disease diagnosis and self-driving cars. In this paper, we propose an effective method to detect adversarial examples in image classification. Our key insight is that adversarial examples are usually sensitive to certain image transformation operations such as rotation and shifting. In contrast, a normal image is generally immune to such operations. We implement this idea of image transformation and evaluate its performance in oblivious attacks. Our experiments with two datasets show that our technique can detect nearly 99% of adversarial examples generated by the state-of-the-art algorithm. In addition to oblivious attacks, we consider the case of white-box attacks. We propose to introduce randomness in the process of image transformation, which can achieve a detection ratio of around 70%.

Recommender system is an important component of many web services to help users locate items that match their interests. Several studies showed that recommender systems are vulnerable to poisoning attacks, in which an attacker injects fake data to a recommender system such that the system makes recommendations as the attacker desires. However, these poisoning attacks are either agnostic to recommendation algorithms or optimized to recommender systems (e.g., association-rule-based or matrix-factorization-based recommender systems) that are not graph-based. Like association-rule-based and matrix-factorization-based recommender systems, graph-based recommender system is also deployed in practice, e.g., eBay, Huawei App Store (a big app store in China). However, how to design optimized poisoning attacks for graph-based recommender systems is still an open problem.

We introduce a new problem, namely, check-in time prediction where the goal is to predict the time when a given user will check-in to a location of interest. We design a novel Recurrent Spatio-Temporal Point Process (RSTPP) model for check-in time prediction. RSTPP addresses two key challenges: 1) Data scarcity due to uneven distribution of check-ins among users/locations. 2) User trajectories contain valuable information that is ignored by standard temporal point process which only considers historical event times. RSTPP is designed to learn the latent dependencies of event times over both historical events and spatio-temporal information about locations a user visited before check-in to the location of interest. We evaluate RSTPP on several real-world datasets, and it significantly outperforms state-of-the-art event time predicting techniques. Our work derives a set of practical implications that can benefit a wide spectrum of applications.

We introduce a novel check-in time prediction problem. The goal is to predict the time a user will check-in to a given location. We formulate check-in prediction as a survival analysis problem and propose a Recurrent-Censored Regression (RCR) model. We address the key challenge of check-in data scarcity, which is due to the uneven distribution of check-ins among users/locations. Our idea is to enrich the check-in data with potential visitors, i.e., users who have not visited the location before but are likely to do so. RCR uses recurrent neural network to learn latent representations from historical check-ins of both actual and potential visitors, which is then incorporated with censored regression to make predictions. Experiments show RCR outperforms state-of-the-art event time prediction techniques on real-world datasets.

We introduce a new query primitive called Function Query (FQ). An FQ operates on a set of math functions and retrieves the functions whose output with a given input satisfies a query condition (e.g., being among top k, within a given range). While FQ finds its natural uses in querying a database of math functions, it can also be applied on a database of discrete values. We show that by interpreting the database as a set of user-defined functions, FQ can achieve the same functionality as existing analytic queries such as top-k query and scalar product query. We address the challenge of efficient execution of FQ. The core of our solution is a novel data structure called Intersection-tree. Our research takes advantage of the fact that 1) the intersections of a set of continuous functions partition their domain into a number of subdomains, and 2) in each of these subdomains, the functions can be sorted based on their output. We evaluate the performance of the proposed techniques through analysis, prototyping, and experiments using both synthetic and real-world data. When querying a database of functions, our techniques scale well. When applied on a database of discrete values, our techniques are more versatile and outperform existing techniques in terms of various performance metrics.

Consider a database where each record represents a math function. A third party is in charge of processing queries over this database and we want to provide a mechanism for users to verify the correctness of their query results. Here each query, referred to as a Function Query (FQ), retrieves the functions whose computation results with user-supplied arguments satisfy certain conditions (e.g., within a certain range). We present authentication solutions that work on a variety of functions, including univariate linear function, multivariate linear function, and multivariate high degree function. Our solutions are based on the fact that the functions can be sorted in the subdomains defined by their intersections and thus can be chained to produce a signature mesh for query result verification. We study the performance of the proposed techniques through theoretical analysis, simulation and empirical study, and include the results in this paper.

Application of stereo video in TV industry and consumer electronics becomes very popular recently. Thus, fast algorithm for stereo video coding is highly desired because of its huge inter-view computational redundancy. In our previous work we proposed an epipolar constraint based fast inter mode selection for stereo video using motion vector of blocks as a indicator of similarity, nevertheless, intra mode selection is also highly complex in standards like H.264/AVC. In this paper, a practical fast intra mode selection is proposed to eliminate the computational redundancy by exploiting inter-view dependency based on epipolar constraint. The proposed method does not rely on disparity estimation. Instead, a sliding window is employed to generate an intra mode candidate pool from macro-blocks on the epipolar line. The candidate pool is then rectified to remove invalid modes and improve accuracy. Finally, optimal prediction mode is selected from the candidate pool. The proposed method can significantly reduce the number of mode candidates/prediction directions compared to exhaustive mode selection by 79%. Experiments on 5 HD video coded in 1-frame demonstrate the overall coding time of one view is saved by 56% on average, with slightly video quality loss less than 0.1 dB.

Fast stereoscopic video encoding becomes a highly desired technique because the stereoscopic video has been realizable for applications like TV broadcasting and consumer electronics. The stereoscopic video has high inter-view dependency subject to epipolar restriction, which can be used to reduce the encoding complexity. In this paper, we propose a fast inter-prediction mode selection algorithm for stereoscopic video encoding. Different from methods using disparity estimation, candidate modes are generated by sliding a window along the macro-block line restricted by the epipolar. Then the motion information is utilized to rectify the candidate modes. A selection failure handling algorithm is also proposed to preserve coding quality. The proposed algorithm is evaluated using independent H.264/AVC encoders for left and right views and can be extended to MVC. Experimental results show that encoding times of one view are reduced by 41.4% and 24.4% for HD and VGA videos respectively with little quality loss.