Bo Ding (丁博)
I am a second-year Ph.D. student in Electrical Engineering at the City University of Hong Kong (CityU), supervised by Professor Haoliang Li. Before CityU, I received my master's degree (2021-2024) from the Institute of Computing Technology, Chinese Academy of Sciences (ICT, CAS), and my bachelor's degree (2017-2021) from Beijing University of Posts and Telecommunications (BUPT).
My research focuses on AI for social media security, with particular emphasis on social bot detection, source identification, and AI-generated content detection. Recently, I have been investigating the implications of LLMs on the security landscape of social media platforms.
Please don't hesitate to contact me if you are interested in my research.
Contact: bo.ding [AT] my.cityu.edu.hk
News
Publications
Keyang Zhang, Chenqi Kong, Hui Liu, Bo Ding, Xinghao Jiang, and Haoliang Li
IEEE Transactions on Information Forensics and Security, 2026
Paper
TL;DR: We propose a Propose-Rectify framework that synergistically combines MLLM's semantic reasoning with multi-scale forensic feature analysis to bridge the gap between high-level understanding and low-level artifact detection for precise image manipulation localization.
Bo Ding*, Tiexin Qin*, Renjie Wan, and Haoliang Li
IEEE Transactions on Knowledge and Data Engineering, 2026
Paper
TL;DR: We propose VoVAE, a probabilistic framework that disentangles dynamic variations from static source features to extract robust representations for generalizable source identification in sequential data.
Bo Ding, Zhenfeng Fan, Zejun Zhao, and Shihong Xia
Multimedia Tools and Applications, 2024
Paper
TL;DR: We propose a multi-branch spatio-temporal difference network for face forgery detection by capturing complementary low-level spatio-temporal features in videos, which can enhance the generalization ability of the model.
Bo Ding, Zhenfeng Fan, Shuang Yang, and Shihong Xia
ArXiv, 2023
Paper
TL;DR: We present Myportrait, a method for learning personalized neural portraits. By incorporating personalized prior from a monocular video and morphable prior from 3DMM, our framework can generate realistic portraits with personalized details.
Zejun Zhao, Zhenfeng Fan, Bo Ding, and Shihong Xia
Frontiers of Data and Computing, 2023
Paper
TL;DR: We design a face forgery detection system based on incremental learning, which reduces the training cost of the model when introducing new forgery samples.
Junhao Dong, Bo Xiao, Bo Ding, and Haoyv Wang
IEEE Access, 2020
Paper
TL;DR: We propose a general transductive ZSL method based on GANs, called GT-GAN, which introduces unlabeled unseen samples for training, to improve the performance of the generator.
Education
University of Chinese Academy of Sciences
Cultivation Unit: Institute of Computing Technology, Chinese Academy of Sciences
School of Information and Communication Engineering
Beijing University of Posts and Telecommunications