About me
I am an incoming master's student at the Institute of Software, Chinese Academy of Sciences. I received my B.Eng. in Information Security from Wuhan University. My work has involved multimodal learning, large models, and intelligent agent systems.
Publications
Dynamic Emotion and Personality Profiling for Multimodal Deception Detection
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026 Main Conference)
We present DDEP, a multimodal dataset with sample-level emotion and personality annotations, and Rel-DDEP, an uncertainty-aware, reliability-weighted fusion framework for joint detection.
Projects
BrowserAide: A Multimodal Large Language Model–Based Web Agent
Project Lead
Developed a browser agent based on a perception–planning–action loop, integrating a FastAPI backend, Chrome extension, and Playwright executor. The system supports DOM, visual, and hybrid control for multi-step web tasks.
github.com/lnennnn/BrowserAide ↗PPTAgent: Generating and Evaluating Presentations Beyond Text-to-Slides
Module Lead
Work on interactive planning, structured outline generation, automated evaluation, and coordinated subagent scheduling for higher-quality and more coherent presentation generation.
github.com/icip-cas/PPTAgent ↗Experience
Large Model Algorithm Intern
Multimodal Big Five Personality Recognition from Video
Built an end-to-end multimodal video pipeline for Big Five personality prediction, covering WhisperX-based preprocessing, automatic annotation, and LoRA/full-parameter fine-tuning of Qwen Omni models. Improved ACC by 0.24% and CCC by 2.27% over the baseline.