avatar

Qian Yang

杨倩

Computer Science @ Mila & Université de Montréal
qian.yang (at) mila.quebec
Quebec, CA


📣 I am on the job market! I am looking for full-time industry positions in Multimodal LLMs and post-training (reward modeling, RL), expected to graduate in Summer / Fall 2027. Please feel free to reach out at qian.yang@mila.quebec!

I am a final-year Ph.D candidate (2023.09 – ) at Mila and Université de Montréal, advised by Prof. Aishwarya Agrawal. My research centers on reliable multimodal LLMs and agents, with a focus on post-training: reward modeling and reinforcement learning. Most recently, at Tencent Hunyuan, I built reward models for text-to-image RL, including a GRPO-trained rubric verifier deployed as the reward service and style-sensitive preference reward models. During my Ph.D I have worked on visual thinking for spatial reasoning in unified multimodal models (EMNLP 2026), online data selection for MLLM training (CVPR 2026), modality alignment for resource-efficient MLLMs (CVPR 2025 Highlight), fine-grained reward modeling for multi-agent collaboration (NeurIPS 2024 Workshop), benchmarking code hallucination (AAAI 2025), and task decomposition for faithful multimodal reasoning (EMNLP 2024). I co-organized the VLMs4All workshop at CVPR 2025, centered on diversity in vision-language models. Before that I obtained my Master's degree in Computer Science from Harbin Institute of Technology, Shenzhen, China (2020-2023) under the supervision of Prof. Baotian Hu. I obtained my Bachelor's degree in Computer Science from the University of Electronic Science and Technology of China (2016-2020). I have been a research intern at Tencent Hunyuan (Qingyun Elite Internship Program) and Alibaba DAMO Academy.

News

Publications ( * denotes equal contribution)

    2026

  1. How and What to Imagine? Visual Thinking in Unified Multimodal Models for Cross-View Spatial Reasoning
    Qian Yang, Ankur Sikarwar, Huy Le, Le Zhang, Zhuan Shi, Perouz Taslakian, Aishwarya Agrawal
    EMNLP 2026.
    EMNLP Code Dataset Project Website
  2. Learning What Matters: Prioritized Concept Learning via Relative Error-driven Sample Selection
    Qian Yang*,Shivam Chandhok*, Oscar Manas, Kanishk Jain, Leonid Sigal, Aishwarya Agrawal
    The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2026.
    CVPR Project Website
  3. Discovering Failure Modes in Vision-Language Models using RL
    Kanishk Jain, Qian Yang, Shravan Nayak, Parisa Kordjamshidi, Nishanth Anand, Aishwarya Agrawal
    Preprint.
    arXiv

    4. 2025

  4. Assessing and Learning Alignment of Unimodal Vision and Language Models
    Le Zhang, Qian Yang, Aishwarya Agrawal
    The IEEE/CVF Conference on Computer Vision and Pattern Recognition 2025.
    CVPR Highlight Project Website

    5. 2024

  5. Decompose and Compare Consistency: Measuring VLMs' Answer Reliability via Task-Decomposition Consistency Comparison
    Qian Yang, Weixiang Yan, Aishwarya Agrawal
    Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing.
    EMNLP
  6. Enhancing Multi-Agent Multi-Modal Collaboration with Fine-Grained Reward Modeling
    Qian Yang, Weixiang Yan, Aishwarya Agrawal
    NeurIPS 2024 Workshop on Adaptive Foundation Models (AFM).
    NeurIPS Workshop
  7. Exploring the Best Practices of Query Expansion with Large Language Models
    Le Zhang, Yihong Wu, Qian Yang, Jian-Yun Nie
    Findings of the Association for Computational Linguistics: EMNLP 2024.
    EMNLP Findings
  8. CodeHalu: Investigating Code Hallucinations in LLMs via Execution-based Verification
    Yuchen Tian*, Weixiang Yan*, Qian Yang, Xuandong Zhao, Qian Chen, Wen Wang, Ziyang Luo, Lei Ma, Dawn Song
    The 39th Annual AAAI Conference on Artificial Intelligence.
    AAAI

    9. 2023

  9. Enhancing Multi-modal and Multi-hop Question Answering via Structured Knowledge and Unified Retrieval-Generation
    Qian Yang, Qian Chen, Wen Wang, Baotian Hu, Min Zhang
    Proceedings of the 31st ACM International Conference on Multimedia, 2023.
    ACM MM

    10. 2022

  10. Chunk-aware Alignment and Lexical Constraint for Visual Entailment with Natural Language Explanations
    Qian Yang*, Yunxin Li*, Baotian Hu, Lin Ma, Yuxing Ding, Min Zhang
    Proceedings of the 30th ACM International Conference on Multimedia, 2022.
    ACM MM
  11. Fast and Robust Online Handwritten Chinese Character Recognition with Deep Spatial & Contextual Information Fusion Network
    Yunxin Li, Qian Yang, Qingcai Chen, Baotian Hu, Xiaolong Wang, Yuxin Ding, Lin Ma
    IEEE Transactions on Multimedia, 2022.
    IEEE Transactions on Multimedia

Awards

Services & Teaching

Education

Internships