Publications

A collection of my research work.

Show, Don't Tell: Morphing Latent Reasoning into Image Generation

Xinxiang Yin, Haodong Chen, Wen-Jie Shu, Hongfei Zhang, Zixin Zhang, Litao Guo, Qifeng Chen, Ying-Cong Chen

International Conference on Machine Learning 2026

A novel framework that seamlessly integrates implicit latent reasoning into the T2I generation process, enhancing Janus-Pro by 16% on GenEval while reducing inference time by 44%.

Text-to-ImageLatent ReasoningImage GenerationVisual Reasoning
arXiv

CodeHacker: Automated Test Case Generation for Detecting Vulnerabilities in Competitive Programming Solutions

Jingwei Shi, Xinxiang Yin, Jing Huang, Jinman Zhao, Shengyu Tao

Annual Meeting of the Association for Computational Linguistics Main Conference 2026

Accepted at the Annual Meeting of the Association for Computational Linguistics 2026 Main Conference. An automated agent framework dedicated to generating targeted adversarial test cases that expose latent vulnerabilities in program submissions, improving the True Negative Rate of existing datasets.

LLM EvaluationCode GenerationTest Case GenerationCompetitive Programming
arXiv

FineQuest: Adaptive Knowledge-Assisted Sports Video Understanding via Agent-of-Thoughts Reasoning

Haodong Chen, Haojian Huang, Xinxiang Yin, Dian Shao

ACM International Conference on Multimedia Oral Presentation 2025

Oral presentation at the ACM International Conference on Multimedia 2025. The first training-free framework for sports video understanding that leverages dual-mode reasoning inspired by cognitive science.

VideoQASports VideoAgent-of-ThoughtsMulti-Modal Learning
arXiv