Biblio

Reward functions are critical hyperparameters with commercial values for individual or distributed reinforcement learning (RL), as slightly different reward functions result in significantly different performance. However, existing inverse reinforcement learning (IRL) methods can be utilized to approximate reward functions just based on collected expert trajectories through observing. Thus, in the real RL process, how to generate a polluted trajectory and perform an adversarial attack on IRL for protecting reward functions has become the key issue. Meanwhile, considering the actual RL cost, generated adversarial trajectories should be minimal and non-catastrophic for ensuring normal RL performance. In this work, we propose a novel approach to craft adversarial trajectories disguised as expert ones, for decreasing the IRL performance and realize the anti-IRL ability. Firstly, we design a reward clustering-based metric to integrate both advantages of fine- and coarse-grained IRL assessment, including expected value difference (EVD) and mean reward loss (MRL). Further, based on such metric, we explore an adversarial attack based on agglomerative nesting algorithm (AGNES) clustering and determine targeted states as starting states for reward perturbation. Then we employ the intrinsic fear model to predict the probability of imminent catastrophe, supporting to generate non-catastrophic adversarial trajectories. Extensive experiments of 7 state-of-the-art IRL algorithms are implemented on the Object World benchmark, demonstrating the capability of our proposed approach in (a) decreasing the IRL performance and (b) having minimal and non-catastrophic adversarial trajectories.

This paper focuses on the typical business scenario of intelligent factory, it includes the manufacturing process, carries out hierarchical security protection, forms a full coverage industrial control security protection network, completes multi-means industrial control security direct protection, at the same time, it utilizes big data analysis, dynamically analyzes the network security situation, completes security early warning, realizes indirect protection, and finally builds a self sensing and self-adjusting industrial network security protection system It provides a reliable reference for the development of intelligent manufacturing industry.

Android malware family classification is an advanced task in Android malware analysis, detection and forensics. Existing methods and models have achieved a certain success for Android malware detection, but the accuracy and the efficiency are still not up to the expectation, especially in the context of multiple class classification with imbalanced training data. To address those challenges, we propose an Android malware family classification model by analyzing the code's specific semantic information based on sensitive opcode sequence. In this work, we construct a sensitive semantic feature-sensitive opcode sequence using opcodes, sensitive APIs, STRs and actions, and propose to analyze the code's specific semantic information, generate a semantic related vector for Android malware family classification based on this feature. Besides, aiming at the families with minority, we adopt an oversampling technique based on the sensitive opcode sequence. Finally, we evaluate our method on Drebin dataset, and select the top 40 malware families for experiments. The experimental results show that the Total Accuracy and Average AUC (Area Under Curve, AUC) reach 99.50% and 98.86% with 45. 17s per Android malware, and even if the number of malware families increases, these results remain good.

Trajectory data in the Internet of Things contains many behavioral information of users, and the method of Mix-zone can be used to separate the association among the user's movement trajectories. In this paper, the weighted undirected graph is used to establish a mathematical model for the Mix-zone, and a user flow-based algorithm is proposed to estimate the probability of migration between nodes in the graph. In response to the attack method basing on the migration probability, the traditional Mix-zone is improved. Finally, an algorithms for adaptively building enhanced Mix-zone is proposed and the simulation using real data sets shows the superiority of the algorithm.

In an augmented reality system, labelling technique is a very useful assistant technique for browsing and understanding unfamiliar objects or environments, through which the superimposed virtual labels of words or pictures on the real scene provide convenient information to the viewers, expand the recognition to area of interests and promote the interaction with real scene. How to design the layout of labels in user's field of view, keep the clarity of virtual information and balance the ratio between virtual information and real scene information is a key problem in the field of view management. This paper presents the empirical results extracted from experiment aiming at the user's visual perception to labelling layout, which reflects the subjective preferences to different factors influencing the labelling result. Statistical analysis of the experiment results shows the intuitive visual judgement accomplished by subjects. The quantitative measurement of clutter indicates the change induced by labels on real scene, therefore contributing the label design on view management in future.