Kaisi Guan (关开思)

profile.png

I am a Master student at Gaoling School of Artificial Intelligence (GSAI), Renmin University of China (RUC). I am advised by Prof. Ruihua Song. Prior to this, I got my bachelor’s degree from GSAI in 2024.

My research centers on Omni understanding & generation, building systems that perceive the world across modalities, and learn to generate it in turn with multimodal.

Main Research Interests:

  • Omni Generation: Building controllable, high-fidelity generative models for video and audio, exploring how their modeling paradigms converge toward a unified framework, along with post-training methods for better alignment and controllability.

  • Omni Understanding: Investigating vision–language–audio interplay and building omni-modal models with stronger understanding of video and audio.

I will graduate in 2027 and am seeking job opportunities in omni generation and understanding. Feel free to contact me at guankaisi@ruc.edu.cn.

News

Sep, 2026 Our paper AVOC was accepted to NeurIPS 2026!
June, 2026 Our papers BridgeDiT and ChronusOmni were accepted to ECCV 2026!
April, 2026 Our paper HuM-Eval was accepted to ICME 2026!
June, 2025 Our paper ETVA was accepted to ICCV 2025!
Sep, 2024 Our paper BSharedRAG was accepted to EMNLP 2024!

Experiences

2026.10 - Now KlingAI @ KlingAI logo Intern
2026.4 - 2026.10 Weixin weixin PNG Image @ , Tencent. Advised by Wenjing Wang and Chen Li Intern
2025.1 - 2025.10 Apple Foundation Model @ Apple PNG Image Apple. Advised by Jeff Lai, Kieran Liu and Meng Cao Intern
2023.1 - 2023.7 ModelBest @ Apple PNG Image Intern
2024.9 - Present AIMind Lab @ Gaoling School of Artificial Intelligence, ruc_logo PNG Image Advised by Ruihua Song Master Student
2020.9 - 2024.06 Gaoling School of Artificial Intelligence, ruc_logo PNG Image RUC Undergraduate student

Publications

  1. arXiv
    SyncReward: Human-Aligned Synchronization Reward for Audio-Visual Generation
    Kaisi Guan, Ying Ba, Wenjing Wang, Xihua Wang, Chen Li, Jing Lyu, Ruihua Song
    arXiv preprint arXiv:2610.16338 2026
    arXiv PDF SyncReward GitHub Repo stars
  2. arXiv
    FATE: Frame-Level Audio-Visual Temporal Embedding
    Kaisi Guan, Bingzi Zhang, Xihua Wang, Ying Ba, Xin Cheng, Yijing Chen, Ruihua Song
    arXiv preprint arXiv:2608.01310 2026
    arXiv PDF FATE GitHub Repo stars
  3. ECCV 2026
    Taming Text-to-Sounding Video Generation via Advanced Modality Condition and Interaction
    Kaisi Guan, Xihua Wang, Zhengfeng Lai, Xin Cheng, Peng Zhang, XiaoJiang Liu, Ruihua Song, Meng Cao
    European Conference on Computer Vision (ECCV 2026).
    arXiv PDF Project Page BridgeDiT GitHub Repo stars
  4. ICCV 2025
    ETVA: Evaluation of Text-to-Video Alignment via Fine-grained Question Generation and Answering
    Kaisi Guan, Zhengfeng Lai, Yuchong Sun, Peng Zhang, Wei Liu, Kieran Liu, Meng Cao, and Ruihua Song
    International Conference on Computer Vision, ICCV 2025, Oct 2025
    arXiv PDF Project Page GitHub Repo stars
  5. EMNLP 2024
    BSharedRAG: Backbone Shared Retrieval-Augmented Generation for the E-commerce Domain
    Kaisi Guan, Qian Cao, Yuchong Sun, Xiting Wang, and Ruihua Song
    In Findings of the Association for Computational Linguistics: EMNLP 2024, Nov 2024
    arXiv PDF Project Page GitHub Repo stars
  6. NeurIPS 2026
    AVOC: Enhancing Hour-Level Audio-Video Understanding in Omni-Modal LLMs via Retrieval-Inspired Token Compression
    Yijing Chen, Wenhui Tan, Xiaoyi Yu, Yuyue Wang, Xin Cheng, Kaisi Guan, Hao Jiang, Xiangyang Li, Guojie Zhu, Ruihua Song
    The Fortieth Annual Conference on Neural Information Processing Systems (NeurIPS 2026).
    arXiv PDF
  7. arXiv
    SyncDPO: Enhancing Temporal Synchronization in Video-Audio Joint Generation via Preference Learning
    Xin Cheng, Xihua Wang, Ying Ba, Yuyue Wang, Kaisi Guan, Yinbo Wang, Wenpu Li, Ruihua Song
    arXiv preprint arXiv:2605.12179 2026
    arXiv PDF Project Page GitHub Repo stars
  8. ICME 2026
    HuM-Eval: A Coarse-to-Fine Framework for Human-Centric Video Evaluation
    Bingzi Zhang, Kaisi Guan, Ruihua Song
    IEEE International Conference on Multimedia and Expo (ICME 2026).
    arXiv PDF
  9. arXiv
    VSSFlow: Unifying Video-conditioned Sound and Speech Generation via Joint Learning
    Xin Cheng, Yuyue Wang, Xihua Wang, Yihan Wu, Kaisi Guan, Yijing Chen, Peng Zhang, Xiaojiang Liu, Meng Cao, Ruihua Song
    arXiv preprint arXiv:2509.24773 2025
    arXiv PDF Project Page GitHub Repo stars
  10. ECCV 2026
    ChronusOmni: Improving Time Awareness of Omni Large Language Models
    Yijing Chen, Yihan Wu, Kaisi Guan, Yuchen Ren, Yuyue Wang, Ruihua Song, Liyun Ru
    European Conference on Computer Vision (ECCV 2026).
    arXiv PDF GitHub Repo stars