Buhua Liu

I am currently an Algorithm Engineer at Hithink RoyalFlush (AInvest), focusing on Financial AI Agent and Reinforcement Learning.

I received my MPhil degree from Hong Kong Baptist University (advised by Prof. Yiu-ming Cheung) and B.Eng. degree from Huazhong University of Science and Technology.

My research interests include:

🤖

LLM Agent

📈

Reinforcement Learning

💹

Financial AI

News

2026.3
Our paper on Zero-Shot Generative Classification is accepted at ICME 2026 (CCF B).
2026.2
Our survey on Diffusion Model Alignment is accepted by ACM Computing Surveys (IF: 28.0)!
2026.1
Our paper on LLM Jailbreak Attack is accepted at ICLR 2026 (CCF A).
2025.2
Our paper on Backdoor Detection is accepted at PAKDD 2025 (Oral).

Selected Publications

Alignment of Diffusion Models: Fundamentals, Challenges, and Future

Buhua Liu, Shitong Shao, Bao Li, Lichen Bai, Zhiqiang Xu, Haoyi Xiong, James Kwok, Sumi Helal, Zeke Xie

ACM Computing Surveys (CSUR), 2026

ACM CSUR IF: 28.0

Ghost in the Cloud: Your Geo-Distributed Large Language Models Training is Easily Manipulated

Zichen Tang*, Zhenheng Tang*, Gaoning Pan, Buhua Liu, Xin He, Kunfeng Lai, Xiaowen Chu, Bo Li

International Conference on Learning Representations (ICLR), 2026

ICLR 2026 CCF A

A Simple and Efficient Baseline for Zero-Shot Generative Classification

Zipeng Qi*, Buhua Liu*, Shiyan Zhang, Bao Li, Zhiqiang Xu, Haoyi Xiong, Zeke Xie

IEEE International Conference on Multimedia and Expo (ICME), 2026

ICME 2026 CCF B

Stronger Separability, Stronger Defense: Influence-based Backdoor Detection

Buhua Liu, Shuo Yang, Zhiqiang Xu, Haoyi Xiong, Yiu-ming Cheung, Zeke Xie

Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD), 2025

PAKDD 2025 CCF C Oral

View All Publications →