Changlian Ma

Changlian Ma

M.S. Student in Computer Science & Technology, Nanjing University

I am a Master's student at Nanjing University, advised by Prof. Limin Wang. I am currently a research intern at Shanghai AI Laboratory (OpenGVLab). My research focuses on Multimodal Large Language Models and Reinforcement Learning for LLMs, with interests in AI4Gaming and AI4Finance. Previously, I received my B.Eng. from Beijing University of Posts and Telecommunications in 2025.

Research Interests

Multimodal Large Language Models Reinforcement Learning for LLMs AI4Gaming AI4Finance

Publications

TimeLens2 overview across seven temporal-grounding benchmarks and query types.

TimeLens2: Generalist Video Temporal Grounding with Multimodal LLMs

Yuhan Zhu, Changlian Ma, Xiangyu Zeng, Xinhao Li, Zhiqiu Zhang, Songze Li, Jun Zhang, Tianxiang Jiang, Yuandong Yang, Ziang Yan, et al.

Show 5 more authors

Zikang Wang, Xinyu Chen, Haoran Chen, Shaowei Zhang, Limin Wang

arXiv 2026

Generalist video temporal grounding with verified multi-span supervision and temporal Wasserstein rewards.

Architecture showing streaming perception, hierarchical caption memory, and proactive responses in OneStreamer.

OneStreamer: Unifying Perception, Memory, and Proactive Response in Streaming Video Interaction

Xiangyu Zeng, Yuandong Yang, Zhiqiu Zhang, Yuhan Zhu, Xinhao Li, Qingyi Si, Dingyu Yao, Changlian Ma, Haoran Chen, Xinyu Chen, et al.

Show 14 more authors

Yansong Shi, Junhao Zhou, Yifei Li, Jun Zhang, Chuanyu Qin, Chenxu Yang, Xinlei Yu, Kun Ouyang, Yuchen Shao, Qianshan Wei, Changhai Zhou, Jun Gao, Jiaqi Wang, Limin Wang

arXiv 2026

Streaming video interaction combining real-time perception, hierarchical caption memory, and evidence-grounded proactive responses.

VideoChat3 overview of temporal perception, long-video understanding, grounding, and online proactive responses.

VideoChat3: Fully Open Video MLLM for Efficient and Generalist Video Understanding

Xinhao Li, Yuhan Zhu, Xiangyu Zeng, Yuhao Dong, Haoning Wu, Zhiqiu Zhang, Yuandong Yang, Changlian Ma, Qingyu Zhang, Yansong Shi, et al.

Show 17 more authors

Xinyu Chen, Haoran Chen, Zizheng Huang, Jun Zhang, Kun Ouyang, Lin Sui, Ziang Yan, Yicheng Xu, Chenting Wang, Yinan He, Hongjie Zhang, Yi Wang, Yu Qiao, Yali Wang, Ziwei Liu, Kai Chen, Limin Wang

arXiv 2026

A fully open 4B video MLLM for efficient general, long-form, and streaming video understanding.

Video-o3 overview illustrating iterative clue seeking with native video tool calls.

Video-o3: Native Interleaved Clue Seeking for Long Video Multi-Hop Reasoning

Xiangyu Zeng, Zhiqiu Zhang, Yuhan Zhu, Xinhao Li, Zikang Wang, Changlian Ma, Qingyu Zhang, Zizheng Huang, Kun Ouyang, Tianxiang Jiang, et al.

Show 5 more authors

Ziang Yan, Yi Wang, Hongjie Zhang, Yali Wang, Limin Wang

ICML 2026

Native video tool use and trajectory-guided reinforcement learning enable iterative clue seeking and long-video multi-hop reasoning.

InternVideo2.5 overview with video-understanding results and fine-grained temporal perception examples.

InternVideo2.5: Empowering Video MLLMs with Long and Rich Context Modeling

Yi Wang, Xinhao Li, Ziang Yan, Yinan He, Jiashuo Yu, Xiangyu Zeng, Chenting Wang, Changlian Ma, Haian Huang, Jianfei Gao, et al.

Show 6 more authors

Min Dou, Kai Chen, Wenhai Wang, Yu Qiao, Yali Wang, Limin Wang

Technical Report 2025

Adaptive token compression and preference optimization strengthen fine-grained perception and long-context video understanding.

Education

Nanjing University

M.S. in Computer Science & Technology · Advisor: Prof. Limin Wang

Sep 2025 – Present

Beijing University of Posts and Telecommunications

B.Eng. in Computer Science & Technology · GPA: 3.84 / 4.0 (Top 1%)

Sep 2021 – Jun 2025

Experience

Shanghai AI Laboratory — OpenGVLab

Nov 2025 – Present

Research Intern

Conduct research on video multimodal foundation models, spanning long-context understanding, temporal grounding, and streaming interaction. Contribute to InternVideo2.5, Video-o3, TimeLens2, VideoChat3, and OneStreamer, with a focus on supervised and reinforcement learning, evidence-grounded reasoning, and memory-based proactive response.

Huawei — Noah's Ark Lab

Jan 2025 – Dec 2025

Research Collaboration

Investigated reinforcement learning for parameter-efficient multimodal models, addressing routing collapse and expert underutilization in LoRA-MoE. Developed RO-GRPO, a mechanism-aware reward framework that improves routing behavior and reasoning performance during GRPO fine-tuning. This work was published at ICLR 2026.

Selected Awards

National Scholarship (China) 2022
Huawei Intelligent Base Scholarship 2023
First Prize (Beijing), Chinese Mathematics Competition (CMC) 2023
Second Prize (National), Lanqiao Cup — C/C++ Programming 2023
Third Prize (National), Huawei ICT Competition — AI Track 2024