👨‍💻 Hongbo Sun (孙宏博) [中文版]

I am currently an AGI researcher at the Institute of Artificial Intelligence (TeleAI), China Telecom, where I focus on developing Multimodal Large Language Models (MLLMs) and advancing their applications in downstream domains. As a core contributor, I participated in the development of China Telecom’s multimodal large language model TeleMM and the TeleSearch 2.0 system designed for ubiquitous surveillance. TeleMM ranked first, second, and third, respectively, on the international authoritative benchmarks MMMU, MME, and the domestic authoritative benchmark OpenCompass (2024 overall leaderboard).

I obtained my Ph.D. degree in Computer Applied Technology at Peking University (PKU) in 2024, earning the Excellent Doctoral Dissertation Award of Beijing Society of Image and Graphics (BSIG). I was selected into Young Elite Scientists Sponsorship Program of the Beijing High Innovation Plan in 2025. My research interests include MLLM, multimodal content understanding, and fine-grained visual analysis.

I am actively open to academic collaborations and recruiting Research Interns at TeleAI. Welcome to contact me with your detailed CV!(Email: sunhb3@chinatelecom.cn)

🔥 News

  • 2025.11 – 1 paper accepted by AAAI 2026.
  • 2025.10 – Ranked 3rd in the ICCV 2025 Multimodal Large Language Model Visual Reasoning Localization Challenge MARS2: VG-RS .
  • 2025.09 – Released the open-source reinforcement learning project for video understanding large model  TSPO , ranking top five in VideoMME, LVBench, and MLVU.
  • 2025.07 – Selected into Young Elite Scientists Sponsorship Program of the Beijing High Innovation Plan, 2025.
  • 2024.12 – Developed TeleMM as the core contributor, which ranked 3rd on the OpenCompass 2024 overall leaderboard, surpassing GPT-4o available at that time. It has been widely applied as the foundational model in China Telecom's businesses such as social security, urban governance, and traffic management.
  • 2024.09 – 1 paper accepted by TIP 2024.
  • 2024.08 – Excellent Doctoral Dissertation Award of Beijing Society of Image and Graphics (BSIG), 2024
  • 2024.07 – Joined Institute of Artificial Intelligence (TeleAI), China Telecom, as a AGI researcher, responsible for developing Multimodal Large Language Models and advancing their applications in downstream domains.
  • 2024.07 – Attained a Ph.D. degree in Computer Application Technology at Peking University (PKU).

📝 Selected Publications

TSPO

TSPO: Temporal Sampling Policy Optimization for Long-form Video Language Understanding

Canhui Tang, Zifan Han, Hongbo Sun, Sanping Zhou, Xuchong Zhang, Xin Wei, Ye Yuan, Huayu Zhang, Jinglin Xu and Hao Sun
AAAI Conference on Artificial Intelligence (AAAI), 2026. (CCF A) (Accepted) [Paper] [Code] [Reported by TeleAI]

SIM-OFE

SIM-OFE: Structure Information Mining and Object-aware Feature Enhancement for Fine-Grained Visual Categorization

Hongbo Sun, Xiangteng He, Jinglin Xu and Yuxin Peng
IEEE Transactions on Image Processing (TIP), Vol. 33, pp. 5312–5326, 2024. (CCF A) [Paper]

FineFMPL

FineFMPL: Fine-grained Feature Mining Prompt Learning for Few-Shot Class Incremental Learning

Hongbo Sun, Jiahuan Zhou, Xiangteng He, Jinglin Xu and Yuxin Peng
Proceedings of the 33rd International Joint Conference on Artificial Intelligence (IJCAI), Jeju, South Korea, Aug. 3-9, 2024. (CCF A)[Paper] [Code]

Dual-Modal Adaptive Online Prompting

Dual-Modal Adaptive Online Prompting and Knowledge Retention for Test-Time Adaptation

Zichen Liu, Hongbo Sun, Yuxin Peng and Jiahuan Zhou
Proceedings of the 38th AAAI Conference on Artificial Intelligence (AAAI), Vancouver, Canada, Feb. 20-27, 2024. (CCF A) [Paper]

HCL

HCL: Hierarchical Consistency Learning for Webly Supervised Fine-Grained Recognition

Hongbo Sun, Xiangteng He and Yuxin Peng
IEEE Transactions on Multimedia (TMM), Vol. 26, pp. 5108–5119, 2024. [Paper] [Code] [Reported by CCF-MM]

Fine-Grained Visual Prompt Learning

Fine-Grained Visual Prompt Learning of Vision-Language Models for Image Recognition

Hongbo Sun, Xiangteng He, Jiahuan Zhou and Yuxin Peng
Proceedings of the 31st ACM International Conference on Multimedia (ACM MM), Ottawa, Canada, Oct. 29-Nov. 3, 2023. (CCF A) [Paper]

SIM-Trans

SIM-Trans: Structure Information Modeling Transformer for Fine-grained Visual Categorization

Hongbo Sun, Xiangteng He and Yuxin Peng
Proceedings of the 30th ACM International Conference on Multimedia (ACM MM), Lisbon, Portugal, Oct. 10-14, 2022. (CCF A)(Oral, 5.9%) [Paper][Code]

📖 Education

  • 2019.09-2024.07  Peking University   Computer Applied Technology          Ph.D.
  • 2016.09-2019.01  Tianjin University   Information and Communication Engineering  Master
  • 2012.09-2016.07  Tianjin University   Electronic Information Engineering        Bachelor

🎖 Honors and Awards

  • Selected into Young Elite Scientists Sponsorship Program of the Beijing High Innovation Plan, 2025
  • Excellent Doctoral Dissertation Award of Beijing Society of Image and Graphics (BSIG), 2024
  • First place in the TRECVID Video Instance Search Competition in 2019 and 2020
  • Third Prize in the National Finals of BOE Innovation Challenge, Achieved Special Offer from BOE Innovation Lab, 2016

📝 Academic Services

  • AAAI Program Committee Member
  • Reviewer for top international conferences and journals (CVPR, ICCV, AAAI, IEEE TMM, etc.)