I am a fourth-year Ph.D. student in the Department of Computer Science at University of Maryland, College Park, advised by Prof. Tianyi Zhou. My research interests are data-centric AI, computer vision (CV), and vision-language models (VLMs), especially exploring synthetic data generation and self-improving learning paradigms to improve model training and evaluation.

I am currently interested in self-improving vision-language models — training VLMs to create their own learning signal (on-policy self-distillation, self-evolving question generation) — and agentic visual systems that actively explore and gather evidence rather than answer in one shot.

I got my M.Eng. degree in Industrial Engineering and Operations Research at University of California, Berkeley. Before joining UCB, I obtained my B.Eng. degree in Industrial Engineering at Tsinghua University.

News

• 2026.07Visual Contrastive Self-Distillation is on arXiv — reward- and teacher-free on-policy self-distillation for VLMs. Project page.
• 2026.07V-REX is accepted to ECCV 2026; History-conditioned token pruning for VLN is accepted to IROS 2026.
• 2026.06Self-Evolving Visual Questioner is on arXiv — VLMs that improve by asking themselves better questions.
• 2025.09ColorBench is accepted to NeurIPS 2025; received a NeurIPS 2025 Student Travel Award.

Publications

• VideoResearcher: Self-Improving Tool Design for Long-Video Understanding

  Dingqiang Ye, Dongdi Zhao, Kaishen Wang, Qingqiao Hu, Jingchen Sun, Yijun Liang, Yuqi Jia, Yiqiao Huang, Yunjie Tian, Jiaxing Zhang, Chuanyang Jin, Ke Zhang, Vishal M. Patel, Di Fu

  arXiv 2026 paper

• 
  Where to Look Matters: On-Policy Self-Distillation for Long-Video Understanding

  Kaishen Wang, Dongdi Zhao, Yijun Liang, Dingqiang Ye, Ruibo Chen, Heng Huang, Di Fu

  arXiv 2026 paper

• 
  Self-Supervised Visual On-Policy Distillation

  Yijiang Li, Yijun Liang, Yunjie Tian, Bingyang Wang, Ke Zhang, Zhenfei Yin, Di Fu, Philip Torr, Nuno Vasconcelos

  arXiv 2026 paper code project page

• 
  On-Policy Self-Distillation without Any Supervision

  Yijiang Li, Bingyang Wang, Yijun Liang, Yunjie Tian, Di Fu, Nuno Vasconcelos

  arXiv 2026 paper code project page

• 
  Visual Contrastive Self-Distillation

  Yijun Liang, Yunjie Tian, Yijiang Li, Yuqi Jia, Furong Huang, Tianyi Zhou, Di Fu

  arXiv 2026 paper code project page

• 
  Self-Evolving Visual Questioner

  Yijun Liang, Hengguang Zhou, Ming Li, Lichen Li, Cho-Jui Hsieh, Tianyi Zhou

  arXiv 2026 paper code project page

• LLMs Struggle to Measure What Distinguishes Students of Different Proficiency Levels: A Study of Item Discrimination in Reading Comprehension Assessment

  Han Chen, Ming Li, Chenguang Wang, Yijun Liang, Dawei Zhou, Hong Jiao, Tianyi Zhou

  arXiv 2026 paper

• Do Prompt-Elicited Trajectories Reflect Training-Time Reward Hacking? A Systematic Study on Monitoring Training-Time Reward Hacking in Code Generation

  Lichen Li, Hengguang Zhou, Yijun Liang, Tianyi Zhou, Cho-Jui Hsieh

  arXiv 2026 paper

• 
  V-REX: Benchmarking Exploratory Visual Reasoning via Chain-of-Questions

  Chenrui Fan*, Yijun Liang*, Shweta Bhardwaj*, Kwesi Cobbina, Ming Li, Tianyi Zhou

  ECCV 2026 paper code

• History-Conditioned Spatio-Temporal Visual Token Pruning for Efficient Vision-Language Navigation

  Qitong Wang, Yijun Liang, Ming Li, Tianyi Zhou, Christopher Rasmussen

  IROS 2026 paper

• 
  ColorBench: Can VLMs See and Understand the Colorful World? A Comprehensive Benchmark for Color Perception, Reasoning, and Robustness

  Yijun Liang*, Ming Li*, Chenrui Fan, Ziyue Li, Dang Nguyen, Kwesi Adu Cobbina, Shweta Bhardwaj, Jiuhai Chen, Fuxiao Liu, Tianyi Zhou

  NeurIPS 2025 paper code dataset

• 
  Diffusion Curriculum: Synthetic-to-Real Generative Curriculum Learning via Image-Guided Diffusion

  Yijun Liang*, Shweta Bhardwaj*, Tianyi Zhou

  ICCV 2025 paper code

• Mosaic-IT: Free Compositional Data Augmentation Improves Instruction Tuning

  Ming Li, Pei Chen, Chenguang Wang, Hongyu Zhao, Yijun Liang, Yupeng Hou, Fuxiao Liu, Tianyi Zhou

  ACL 2025 paper

• CaughtCheating: Is Your MLLM a Good Cheating Detective? Exploring the Boundary of Visual Perception and Reasoning

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

  arXiv 2025 paper

• PEDANTS: Cheap but Effective and Interpretable Answer Equivalence

  Zongxia Li, Ishani Mondal, Yijun Liang, Huy Nghiem, Jordan Lee Boyd-Graber

  EMNLP 2024 paper

• GeoDRL: A Self-Learning Framework for Geometry Problem Solving Using Reinforcement Learning in Deductive Reasoning

  Shuai Peng, Di Fu, Yijun Liang, Liangcai Gao, Zhi Tang

  ACL 2023 paper

• Compute Like Humans: Interpretable Step-by-Step Symbolic Computation with Deep Neural Network

  Shuai Peng, Di Fu, Yong Cao, Yijun Liang, Gu Xu, Liangcai Gao, Zhi Tang

  KDD 2022 paper

Browse by topic →

Education

• University of Maryland, College ParkPh.D. in Computer Science
  2023 – present
• University of California, BerkeleyM.Eng. in Industrial Engineering & Operations Research
  2019 – 2020
• Tsinghua UniversityB.Eng. in Industrial Engineering
  2015 – 2019

Awards & Service

• Silver Reviewer Award, ICML 2026
• Student Travel Award, NeurIPS 2025
• Dean's Fellowship & Chair's Fellowship, University of Maryland, 2023
• Reviewer: ICLR 2027, ICML 2026  ·  Program Committee: AAAI 2026