Visible to the public Biblio

Filters: Author is Liu, Tao  [Clear All Filters]
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z 
Zheng, Wei, Abdallah Semasaba, Abubakar Omari, Wu, Xiaoxue, Agyemang, Samuel Akwasi, Liu, Tao, Ge, Yuan.  2021.  Representation vs. Model: What Matters Most for Source Code Vulnerability Detection. 2021 IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER). :647–653.
Vulnerabilities in the source code of software are critical issues in the realm of software engineering. Coping with vulnerabilities in software source code is becoming more challenging due to several aspects of complexity and volume. Deep learning has gained popularity throughout the years as a means of addressing such issues. In this paper, we propose an evaluation of vulnerability detection performance on source code representations and evaluate how Machine Learning (ML) strategies can improve them. The structure of our experiment consists of 3 Deep Neural Networks (DNNs) in conjunction with five different source code representations; Abstract Syntax Trees (ASTs), Code Gadgets (CGs), Semantics-based Vulnerability Candidates (SeVCs), Lexed Code Representations (LCRs), and Composite Code Representations (CCRs). Experimental results show that employing different ML strategies in conjunction with the base model structure influences the performance results to a varying degree. However, ML-based techniques suffer from poor performance on class imbalance handling when used in conjunction with source code representations for software vulnerability detection.
Han, Rui, Du, Liping, Liu, Tao, Chen, Yueyun.  2017.  SVM-GA Based Method for Estimation of a Large Number of Primary Users in Mobile Cognitive Radio Networks. Proceedings of the 3rd International Conference on Communication and Information Processing. :311–315.

In cognitive radio networks with mobile terminals, it is not enough for spectrum sensing only to determine whether primary user (PU) occupy the spectrum band. Sometimes we also want to know more priori information, such as, the number of PUs, which can help to estimate its carrier frequency, direction of arrival, and location. In this paper, a machine learning based method is proposed to estimate a large number of primary users. In the proposed method, support vector machine (SVM) is used to achieve the number of primary users while genetic algorithm (GA) is to optimize the parameters of SVM kernel. The first class feature of SVM is the ratio of the element sum and the trace of sample covariance matrix, and the second class feature is the mean of Gerschgorin radii. The simulation results show that our proposed SVM-GA algorithm has higher accuracy than SVM.

Deng, Yingjie, Zhao, Dingxuan, Liu, Tao.  2021.  Self-Triggered Tracking Control of Underactuated Surface Vessels with Stochastic Noise. 2021 International Conference on Security, Pattern Analysis, and Cybernetics(SPAC). :266–273.
This note studies self-triggered tracking control of underactuated surface vessels considering both unknown model dynamics and stochastic noise, where the measured states in the sensors are intermittently transmitted to the controller decided by the triggering condition. While the multi-layer neural network (NN) serves to approximate the unknown model dynamics, a self-triggered adaptive neural model is fabricated to direct the design of control laws. This setup successfully solves the ``jumps of virtual control laws'' problem, which occurs when combining the event-triggered control (ETC) with the backstepping method, seeing [1]–[4]. Moreover, the adaptive model can act as the filter of states, such that the complicated analysis and control design to eliminate the detrimental influence of stochastic noise is no longer needed. Released from the continuous monitoring of the controller, the devised triggering condition is located in the sensors and designed to meet the requirement of stability. All the estimation errors and the tracking errors are proved to be exponentially mean-square (EMS) bounded. Finally, a numerical experiment is conducted to corroborate the proposed strategy.