I am a fourth-year PhD candidate in Computer Science at the University of Maryland, College Park, in the CLIP Lab, advised by Jordan Boyd-Graber and co-advised by Lichao Sun. My research develops methods that make vision-language models and agents more capable through post-training, more autonomous through self-evolution, and more useful through tighter collaboration with humans.

Research Focus

1. Model Post-Training — Reinforcement learning and reward design for vision-language models, including self-rewarding via reasoning decomposition, semantically-aware open-ended rewards, and hallucination-targeted alignment for image and video understanding.

2. Self-Evolving Agents — Autonomous agents that bootstrap their own capabilities from minimal or zero human-curated data, spanning self-evolving multimodal reasoning, co-evolving decision-and-skill agents for long-horizon tasks, and exploration-guided visual reasoning.

3. Human-AI Collaboration — How humans and AI systems collaborate to optimize real-world workflows, from LLM-assisted annotation and interactive topic exploration to robust automatic evaluation metrics that faithfully reflect human judgment.

News

Selected Publications

An Agentic Harness for Skill-Evolving Image Generation Workflows

Zongxia Li*, Dawei Liu*, Jingxi Chen, Xiyang Wu, Lichao Sun, et al.

Preprint

[paper] [code]

Self-Rewarding Vision-Language Model via Reasoning Decomposition

Zongxia Li*, Wenhao Yu*, Chengsong Huang, Rui Liu, Zhenwen Liang, Fuxiao Liu, Jingxi Chen, Dian Yu, Jordan Boyd-Graber, Haitao Mi, Dong Yu

ICLR 2026

[paper] [code]

MM-Zero: Self-Evolving Multi-Model Vision Language Models From Zero Data

Zongxia Li*, Hongyang Du*, Chengsong Huang*, Xiyang Wu, Lantao Yu, Yicheng He, Jing Xie, et al.

Preprint

[paper] [code]

VideoHallu: Evaluating and Mitigating Multi-modal Hallucinations for Synthetic Videos

Zongxia Li*, Xiyang Wu*, Yubin Qin, Hongyang Du, Guangyao Shi, Dinesh Manocha, Tianyi Zhou, Jordan Lee Boyd-Graber

NeurIPS 2025

[paper] [webpage] [code]

First Frame Is the Place to Go for Video Content Customization

Jingxi Chen*, Zongxia Li*, Zhichao Liu, Guangyao Shi, Xiyang Wu, Fuxiao Liu, Cornelia Fermüller, Brandon Y. Feng, Yiannis Aloimonos

CVPR 2026

[paper] [webpage]

Graph-of-Skills: Dependency-Aware Structural Retrieval for Massive Agent Skills

Dawei Liu*, Zongxia Li*, Hongyang Du, Xiyang Wu, Lichao Sun

Preprint

[paper] [code]

Co-Evolving LLM Decision and Skill Bank Agents for Long-Horizon Tasks

Xiyang Wu, Zongxia Li, Guangyao Shi, Alexander Duffy, Tyler Marques, Matthew Lyle Olson, Tianyi Zhou, Dinesh Manocha

Preprint

[paper] [webpage]

See all publications »

Popular Community Projects