UIT Students' Scientific Paper Accepted at SOICT 2023 International Conference

Wed, 25/10/2023 - 23:11

Paper Title: "Binary Representation Embedding and Deep Learning for Binary Code Similarity Detection in Software Security Domain"

Authors:

Nguyen Hung Thinh - 20521962 - ATCL2020 - Primary Author

Nguyen Phuc Hai - 20521281 - ATTT2020 - Co-Author

Tran Dinh Khoa - 22520689 - ATTN2022 - Co-Author

Le Tran Thanh Nhan - 19520799 - ATTN2019 - Co-Author

Supervisors:

Dr. Pham Van Hau

Ms. Phan The Duy

Abstract:

Binary Code Similarity Detection (BCSD) involves analyzing the binary representation of two functions, programs, or related entities to produce a numerical output indicating their similarity level. BCSD research covers various tasks, including searching for binary patterns, such as finding code snippets within files, or determining binary functions similar to a specific binary code. This capability opens up many potential applications for BCSD, such as detecting software vulnerabilities in a large software set requiring testing, identifying and analyzing malicious software, and detecting unique malware codes. In this paper, we introduce BiSim-Inspector, a binary code similarity detection tool based on deep learning. This tool leverages the "bytes2vec" method developed by us to convert bytecode of binary functions into vectors, which are then fed into a Convolutional Neural Network - Gated Recurrent Unit (CNN-GRU) model to assess their similarity. We conducted a series of experiments to evaluate the proposed method by comparing it with the best tools in the current field. We utilized a large, well-structured, and diverse dataset created for binary code similarity detection tasks, called BinaryCorp. The experimental results show that our model achieved a Recall rate of 89%, which is 25% higher than comparable methods, while maintaining reasonable computational costs and consistent training and prediction times.

We express our sincere gratitude to Professor Pham Van Hau, Head of Information Security, for creating an exceptionally dynamic academic environment for us students to participate in. We also thank Mr. Phan The Duy, Deputy Head of the Information Security Laboratory, for his continuous guidance, direction, and shared enthusiasm for research, enabling us to refine our research work and enhance our skills. Special thanks to Mr. Ngo Khanh Khoa and Mr. To Trong Nghia for their valuable feedback, which helped our team refine the content and achieve the publication of our scientific paper at the SOICT 2023 conference.

The SOICT (Symposium on Information and Communication Technology) is an international conference covering essential research areas such as AI Foundations and Big Data, Network Communication and Security, Image and Natural Language Processing, Software Engineering and Digital Technology, Blockchain, and Operations Research trends. The 12th SOICT 2023 international conference will be held in Ho Chi Minh City, organized by The School of Information and Communications Technology, Hanoi University of Science and Technology, VNUHCM- University of Scienc, Laboratory Informatics, Modelling and Optimisation System (LIMOS), The French National Centre for Scientific Research (CNRS), and the Institute for Advanced Study in Mathematics. The conference, scheduled for December 7-8, 2023, aims to provide an academic forum for postgraduate researchers and students to share the latest research findings and identify future challenges in computer science.