Songyang Zhang

Young Scientist at Shanghai AI Laboratory, Shanghai, China.

zhangsongyang_2024.jpg

I am a Young Scientist of Shanghai AI Laboratory, collaborate with Dr. Kai Chen. Previously, I was a postdoctoral researcher, supervised by Prof. Dahua Lin. Currently, I lead a team working on foundation model, includes the research and open-source platform. My team intiates the OpenCompass project, an evaluation platform for foundation model. We also developed OpenMMLab projects MMPreTrain.

OpenCompass MMPreTrain

I obtained Ph.D. in Computer Science at the University of Chinese Academy of Science(UCAS), in the joint program at PLUS Lab, ShanghaiTech University in 2022. I got my B.Sc. degree in 2017 from Beihang University.

Open positions include full-time researchers/engineers and interns, feel free to contact me through the email. Research directions include: evaluation, alignment of large language model(agent, reasoning, interpretability, etc), multi-modality learning( vision/audio/-language learning), etc.

News

Jun 29, 2025 1 paper on diffusion model for image generation accepted to ICCV 2025.
May 21, 2025 4 papers (2 main + 2 findings) have been accepted to ACL 2025. One of the papers has been selected as an oral presentation.
Dec 28, 2024 I am honored to serve as the Area Chair (AC) for the ACL Rolling Review (ARR) in December 2024.
Dec 9, 2024 We are excited to announce that one of our papers has been accepted for presentation at AAAI 2025!
Sep 27, 2024 Three papers(2 conference papers and 1 DB track paper) are accepted by NeurIPS 2024!
Sep 20, 2024 Two papers(LawBench and ProSA) are accepted by EMNLP 2024!
Jul 20, 2024 I have finished the postdoctral research project and join Shanghai AI Lab as the Young Scientist.
Jul 1, 2024 MMBench is accepted by ECCV 2024! Congratulations to Yuan Liu.
Jun 16, 2024 I was awarded the WAIC 2024 Rising Star (15 AI researchers in total).
May 16, 2024 4 papers (2 main + 2 findings) accepted to ACL 2024.

Selected Publications(Full List)

  1. ACL
    Capability Salience Vector: Fine-grained Alignment of Loss and Capabilities for Downstream Task Scaling Law
    Qiming Ge, Shuhao Xing, Songyang Gao, Yunhua Zhou, Yicheng Zou, and 6 more authors
    In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2025
  2. ACL
    Are Your LLMs Capable of Stable Reasoning?
    Junnan Liu, Hongwei Liu, Linchen Xiao, Ziyi Wang, Kuikun Liu, and 4 more authors
    In Findings of the Association for Computational Linguistics (ACL), 2025
  3. ACL
    Condor: Enhance LLM Alignment with Knowledge-Driven Data Synthesis and Refinement
    Maosong Cao, Taolin Zhang, Mo Li, Chuyu Zhang, Yunxin Liu, and 3 more authors
    In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2025
  4. NeurIPS
    InternLM-XComposer2-4KHD: A Pioneering Large Vision-Language Model Handling Resolutions from 336 Pixels to 4K HD
    Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Bin Wang, and 19 more authors
    In Proceeding of Advances in Neural Information Processing Systems (NeurIPS), 2024
  5. EMNLP
    ProSA: Assessing and Understanding the Prompt Sensitivity of LLMs
    Zhuo Jingming, Zhang Songyang, Fang Xinyu, Duan Haodong, Lin Dahua, and 1 more author
    In Findings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2024
  6. EMNLP
    LawBench: Benchmarking Legal Knowledge of Large Language Models
    Zhiwei Fei, Xiaoyu Shen, Dawei Zhu, Fengzhe Zhou, Zhuo Han, and 4 more authors
    In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), 2024
  7. ACL
    MathBench: Evaluating the Theory and Application Proficiency of LLMs with a Hierarchical Mathematics Benchmark
    Hongwei Liu, Zilong Zheng, Yuxuan Qiao, Haodong Duan, Zhiwei Fei, and 5 more authors
    In Findings of the Association for Computational Linguistics (ACL), 2024
  8. ACL
    T-Eval: Evaluating the Tool Utilization Capability of Large Language Models Step by Step
    Zehui Chen, Weihua Du, Wenwei Zhang, Kuikun Liu, Jiangning Liu, and 6 more authors
    In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), 2024
  9. ArXiv
    InternLM2 Technical Report
    Zheng Cai, Maosong Cao, Haojiong Chen, Kai Chen, Keyu Chen, and 95 more authors
    2024
  10. ArXiv
    InternLM-XComposer2: Mastering Free-form Text-Image Composition and Comprehension in Vision-Language Large Model
    Xiaoyi Dong, Pan Zhang, Yuhang Zang, Yuhang Cao, Bin Wang, and 18 more authors
    2024
  11. CVPR
    From Pixels to Graphs: Open-Vocabulary Scene Graph Generation with Vision-Language Models
    Rongjie Li, Songyang Zhang, Dahua Lin, Kai Chen, and Xuming He
    In Proceeding of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024
  12. T-PAMI
    SGTR+: End-to-end Scene Graph Generation with Transformer
    Rongjie Li, Songyang Zhang, and Xuming He
    In IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2024
  13. ECCV
    MMBench: Is Your Multi-modal Model an All-around Player?
    Yuan Liu, Haodong Duan, Yuanhan Zhang, Bo Li, Songyang Zhang, and 7 more authors
    In Proceeding of the European Conference on Computer Vision (ECCV), 2024
  14. ICCV
    Improving Pixel-based MIM by Reducing Wasted Modeling Capability
    Yuan Liu, Songyang Zhang, Jiacheng Chen, Zhaohui Yu, Kai Chen, and 1 more author
    In Proceedings of the IEEE/CVF International Conference on Computer Vision(ICCV), 2023
  15. CVPR
    RIFormer: Keep Your Vision Backbone Effective But Removing Token Mixer
    Jiahao Wang, Songyang Zhang, Yong Liu, Taiqiang Wu, Yujiu Yang, and 4 more authors
    In Proceeding of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023
  16. NeurIPS
    Dynamic Grained Encoder for Vision Transformers
    Lin Song*, Songyang Zhang*, Songtao Liu, Zeming Li, Xuming He, and 3 more authors
    In Proceeding of Advances in Neural Information Processing Systems (NeurIPS), 2021
  17. CVPR
    Distribution Alignment: A Unified Framework for Long-tail Visual Recognition
    Songyang Zhang, Zeming Li, Shipeng Yan, Xuming He, and Jian Sun
    In Proceeding of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021
  18. ICML
    LatentGNN: Learning Efficient Non-local Relations for Visual Recognition
    Songyang Zhang, Shipeng Yan, and Xuming He
    In Proceeding of the 36th International Conference on Machine Learning (ICML),, 2019