Publications | Xiyang Wu

Preprints and an indexed list are also on my Google Scholar page.

2026

ASSEMBLE: Atomic Skills for Evidence-Grounded Video Reasoning [arXiv]

Xiyang Wu, Zongxia Li, Shengxin Zhang, Zhichao Liu, Dinesh Manocha

arXiv 2026
- Learning
- RL Post-training
- Video Reasoning
- Skill Composition
COS-PLAY: Co-Evolving LLM Decision and Skill Bank Agents for Long-Horizon Tasks [arXiv]

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

arXiv 2026

[Website] [GitHub] [HF Dataset]
- Learning
- RL Post-training
- Skill Learning
- Long-horizon Agents
SABER: A Stealthy Agentic Black-Box Attack Framework for Vision-Language-Action Models [arXiv]

Xiyang Wu, Guangyao Shi, Qingzi Wang, Zongxia Li, Amrit Singh Bedi, Dinesh Manocha

IROS 2026 · Oral Presentation

[Website] [GitHub] [Hugging Face]
- Acting
- VLA Models
- Adversarial Robustness
- Red Teaming
MASS: Motion-Aware Spatial-Temporal Grounding for Physics Reasoning and Comprehension in Vision-Language Models [arXiv]

Xiyang Wu^*, Zongxia Li, Jihui Jin, Guangyao Shi, Gouthaman KV, Vishnu Raj, Nilotpal Sinha, Jingxi Chen, Fan Du, Dinesh Manocha

CVPR 2026
- Understanding
- World Model
- VLM
- Video Understanding
COMFYCLAW: Self-Evolving Skill Harnesses for Image Generation Workflows [arXiv]

Zongxia Li, Dawei Liu, Fuxiao Liu, Yuhang Zhou, Xiyang Wu, Jingxi Chen, Jing Xie, Xiaomin Wu, Lichao Sun

arXiv 2026

[GitHub]
- Learning
- Self-improving Agents
- Agent Skills
- Image Generation
Recursive Agentic Reasoning [arXiv]

Shengxin Zhang, Xiaomin Wu, Xiyang Wu, Jing Xie

arXiv 2026
- Learning
- Agentic Reasoning
- Test-time Scaling
Long-Horizon-Terminal-Bench: Testing the Limits of Agents on Long-Horizon Terminal Tasks with Dense Reward-Based Grading [arXiv]

Zongxia Li, Zhongzhi Li, Yucheng Shi, Ruhan Wang, Junyao Yang, Zhichao Liu, Xiyang Wu, Anhao Li, Yue Yu, Ninghao Liu, Lichao Sun, Haotao Mi, Leowei Liang

arXiv 2026

[Website] [GitHub]
- Learning
- Agent Evaluation
- Long-horizon Agents
- Reward Design
Causeway: Restoring Task Accessibility for Instruction Switching in VLA Policies [arXiv]

Qingzi Wang, Kaixi Feng, Guangyao Shi, Xiyang Wu, Ang Li, Dinesh Manocha

arXiv 2026
- Acting
- VLA Models
- Instruction Following
- Task Recovery
Act on What You See: Unlocking Safe Social Navigation in Vision-Language-Action Models [arXiv]

Qingzi Wang, Xiyang Wu, Guangyao Shi, Dianwei Chen, Xianfeng Yang, Dinesh Manocha

arXiv 2026
- Acting
- VLA Models
- Robot Safety
- Social Navigation
First Frame Is the Place to Go for Video Content Customization [arXiv]

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

CVPR 2026
- Understanding
- Video Generation
- Content Customization
MM-Zero: Self-Evolving Multi-Model Vision Language Models From Zero Data [arXiv]

Zongxia Li, Hongyang Du, Chengsong Huang, Xiyang Wu, Lantao Yu, Yicheng He, Jing Xie, Xiaomin Wu, Zhichao Liu, Jiarui Zhang, Fuxiao Liu

EMNLP 2026
- Learning
- RL Post-training
- Self-improving Models
- Multimodal Reasoning
Graph-of-Skills: Dependency-Aware Structural Retrieval for Massive Agent Skills [arXiv]

Dawei Liu, Zongxia Li, Hongyang Du, Xiyang Wu, Shihang Gui, Yongbei Kuang, Lichao Sun

EMNLP 2026

[GitHub]
- Learning
- Skill Retrieval
- Skill Composition
- Agent Skills
FALCON: Future-Aware Learning with Contextual Object-Centric Pretraining for UAV Action Recognition [arXiv]

Ruiqi Xian, Xiyang Wu, Tianrui Guan, Xijun Wang, Boqing Gong, Dinesh Manocha

IROS 2026
- Acting
- UAV Action Recognition
- Self-supervised Learning
Multimodal Video Generation Models with Audio: Present and Future [Paper]

Zongxia Li, Hongyang Du, Dawei Liu, Xiyang Wu, Lantao Yu, Jingxi Chen, Fuxiao Liu, Xiaomin Wu, Jing Xie, Chengsong Huang, Yicheng He, Guangyao Shi

Preprint 2026

[GitHub]
- Understanding
- Video Generation
- Multimodal Models
- Audio Generation

2025

On the Vulnerability of LLM/VLM-Controlled Robotics [arXiv]

Xiyang Wu, Souradip Chakraborty, Ruiqi Xian, Jing Liang, Tianrui Guan, Fuxiao Liu, Brian Sadler, Dinesh Manocha, Amrit Singh Bedi

IROS 2025

[Website] [GitHub]
- Acting
- Embodied AI
- Adversarial Robustness
- AI Safety
VideoHallu: Evaluating and Mitigating Multi-modal Hallucinations on Synthetic Video Understanding [arXiv]

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

NeurIPS 2025

[Website] [GitHub] [HF Dataset]
- Understanding
- World Model
- VLM
- Video Understanding
CaughtCheating: Is Your MLLM a Good Cheating Detective? Exploring the Boundary of Visual Perception and Reasoning [arXiv]

Ming Li, Chenguang Wang, Yijun Liang, Xiyao Wang, Yuhang Zhou, Xiyang Wu, Yuqing Zhang, Ruiyi Zhang, Tianyi Zhou

arXiv 2025
- Understanding
- Visual Perception
- Multimodal Reasoning
Semantically-Aware Rewards for Open-Ended R1 Training in Free-Form Generation [arXiv]

Zongxia Li, Yapei Chang, Yuhang Zhou, Xiyang Wu, Zichao Liang, Yoo Yeon Sung, Jordan Lee Boyd-Graber

arXiv 2025

[GitHub]
- Learning
- RL Post-training
- Reward Design
- Free-form Generation
A Survey of State of the Art Large Vision Language Models: Alignment, Benchmark, Evaluations and Challenges [arXiv]

Zongxia Li, Xiyang Wu, Hongyang Du, Fuxiao Liu, Huy Nghiem, Guangyao Shi

CVPRW 2025
- Understanding
- Vision-Language Models
- Multimodal Evaluation
LLM-as-a-Judge Failures at Automating the Identification of Poor Quality Outputs in Free-Form Texts [Paper]

Zongxia Li, Xiyang Wu, Ishani Mondal, Alexa Siu, Jordan Boyd-Graber, Ani Nenkova

NewSum 2025

[PDF] [DOI]
- Understanding
- LLM Evaluation
- LLM-as-a-Judge
- Evaluation Reliability
Taming Large Language Models for Free-Form Generation Via Reinforcement Learning With Verifiable Rewards [Paper]

Zongxia Li, Yapei Chang, Yuhang Zhou, Xiyang Wu, Yoo Yeon Sung, Zichao Liang, Jordan Lee Boyd-Graber

Manuscript 2025

[PDF]
- Learning
- RL Post-training
- Verifiable Rewards

2024

AUTOHALLUSION: Automatic Generation of Hallucination Benchmarks for Vision-Language Models [arXiv]

Xiyang Wu^*, Tianrui Guan^*, Dianqi Li, Shuaiyi Huang, Xiaoyu Liu, Xijun Wang, Ruiqi Xian, Abhinav Shrivastava, Furong Huang, Jordan Lee Boyd-Graber, Tianyi Zhou, Dinesh Manocha

EMNLP 2024

[Website] [GitHub]
- Understanding
- Synthetic Data Curation
- Multimodal Evaluation
- Hallucination
LANCAR: Leveraging Language for Context-Aware Robot Locomotion in Unstructured Environments [arXiv]

Xiyang Wu^*, Chak Lam Shek^*, Wesley A. Suttle, Carl Busart, Erin Zaroukian, Dinesh Manocha, Pratap Tokekar, Amrit Singh Bedi

IROS 2024

[Website]
- Acting
- Robot Learning
- Reinforcement Learning
- Language Grounding
HallusionBench: An Advanced Diagnostic Suite for Entangled Language Hallucination and Visual Illusion in Large Vision-Language Models [arXiv]

Tianrui Guan^*, Fuxiao Liu^*, Xiyang Wu, Ruiqi Xian, Zongxia Li, Xiaoyu Liu, Xijun Wang, Lichang Chen, Furong Huang, Yaser Yacoob, Dinesh Manocha, Tianyi Zhou

CVPR 2024

[GitHub]
- Understanding
- Multimodal Evaluation
- Visual Reasoning
- Hallucination
AGL-NET: Aerial-Ground Cross-Modal Global Localization with Varying Scales [arXiv]

Tianrui Guan^*, Ruiqi Xian^*, Xijun Wang, Xiyang Wu, Mohamed Elnoor, Daeun Song, Dinesh Manocha

IROS 2024
- Acting
- Cross-modal Localization
- Aerial-ground Matching

2023

iPLAN: Intent-Aware Planning in Heterogeneous Traffic via Distributed Multi-Agent Reinforcement Learning [arXiv]

Xiyang Wu, Rohan Chandra, Tianrui Guan, Amrit Singh Bedi, Dinesh Manocha

CoRL 2023 Oral

[GitHub]
- Learning
- Multi-agent RL
- Intent-aware Planning
- Autonomous Driving
A novel image registration-based dynamic photometric stereo method for online defect detection in aluminum alloy castings [Paper]

Haoyue Liu, Xiyang Wu, Ning Yan, Zexiao Li, Xiaodong Zhang

DSP 2023
- Acting
- Photometric Stereo
- Defect Detection

2020

FireCommander: An Interactive, Probabilistic Multi-agent Environment for Heterogeneous Robot Teams [arXiv]

Esmaeil Seraj, Xiyang Wu, Matthew Gombolay

arXiv 2020
- Learning
- Multi-agent Systems
- Heterogeneous Robots