I am currently a Ph.D. student at the School of Artificial Intelligence, Beihang University, and a member of CoLab, advised by Prof. Si Liu, with an expected graduation date of 2027.06. I am also a research intern at Shanghai AI Laboratory, where I am supervised by Prof. Jiangmiao Pang and Prof. Jifeng Dai. Prior to my Ph.D., I spent two years in the masterβs program at Beihang University under the supervision of Prof. Lv Sheng, and I received my B.Eng. from Beihang University. My research interests lie in Embodied AI and Multimodal Large Language Models, I also have research interests and experience in Reinforcement Learning.
π₯ News
- π 2026.06: RoboInter++ was released as a world-action-modeling extension of RoboInter for robotic manipulation and embodied world modeling.
- π 2026.05: LLaVA-CMoE was accepted to ICML 2026.
- π 2026.02: Data, benchmarks, and models from RoboInter were open-sourced.
- π 2026.02: As team leader, I led our team to the Second Place Award (2/62) in the RoCo Challenge @ AAAI 2026: Robotic Collaborative Assembling for Human-Centered Manufacturing.
- π 2026.01: RoboInter was accepted to ICLR 2026.
- π 2025.11: VaccineRAG was accepted to AAAI 2026.
- π 2025.09: InternVLA-M1 was released as a technical report.
- π 2025.03: AirStar was accepted to ACM MM 2025.
- π 2025.01: OpenUAV was accepted to ICLR 2025.
- π 2024.09: CooHOI was accepted to NeurIPS 2024 as a Spotlight paper.
- π 2024.03: I joined Shanghai AI Laboratory as a research intern.
- π 2024.01: Octavius was accepted to ICLR 2024.
- π 2023.05: I joined NIO as a research intern.
- π 2023.03: VL-SAT was accepted to CVPR 2023 as a Highlight paper.
- π 2021.09: I joined SenseTime as a research intern.
Publications
Equal-contribution authors are in bold.
RoboInter++: A Holistic Intermediate Representation Suite for Robotic Manipulation and Embodied World Modeling
Ziqin Wang, Hao Li, Weijun Wang, Junhao Cai, Jia Zeng, Jiangmiao Pang, Yilun Chen, Si Liu
Technical Report
LLaVA-CMoE: Towards Continual Mixture of Experts for Large Vision-Language Models
Ziqin Wang, Hengyuan Zhao, Qixin Sun, Kaiyou Song, Yilin Li, Xiaolin Hu, Qingpei Guo, Si Liu
ICML, 2026
VaccineRAG: Boosting Multimodal Large Language Modelsβ Immunity to Harmful RAG Samples
Qixin Sun, Ziqin Wang, Hengyuan Zhao, Yilin Li, Kaiyou Song, Linjiang Huang, Xiaolin Hu, Qingpei Guo, Si Liu
AAAI 2026
Towards Realistic UAV Vision-Language Navigation: Platform, Benchmark, and Methodology
Xiangyu Wang, Ziqin Wang, Donglin Yang, Hohin Kwan, Jinyu Chen, Wenjun Wu, Hongsheng Li, Yue Liao, Si Liu
ICLR 2025
CooHOI: Learning Cooperative Human-Object Interaction with Manipulated Object Dynamics
Ziqin Wang, Jiawei Gao, Zeqi Xiao, Jingbo Wang, Tai Wang, Jinkun Cao, Xiaolin Hu, Si Liu, Jifeng Dai, Jiangmiao Pang
NeurIPS 2024, Spotlight
Octavius: Mitigating Task Interference in MLLMs via LoRA-MoE
Zeren Chen, Ziqin Wang, Zhen Wang, Huayang Liu, Zhenfei Yin, Si Liu, Lu Sheng, Wanli Ouyang, Yu Qiao, Jing Shao
ICLR 2024
Research and Industry Experience
Worked on multi-agent collaboration, robotic manipulation, embodied world models, and VLA systems. Contributed to InternVLA-M1, RoboInter, and RoboInter++ related projects, including a co-first-author NeurIPS Spotlight paper.
Worked on point-cloud perception foundation models for autonomous driving and improved cone detection recall while optimizing inference speed for deployment.
Participated in InternVL pretraining and CLIP reproduction and optimization, and explored multimodal large models with mixture-of-experts architectures.
Worked on liveness detection algorithms, covering data collection, model iteration, and practical system improvement for real-world deployment.
Education
Ph.D. student, School of Artificial Intelligence.
M.Eng., School of Software.
B.Eng., School of Software.
Projects
AirStar
LLM-Agent UAV Assistant
An LLM-agent-based UAV assistant system that integrates skill selection and tool calling to support campus navigation, photo capture, following, and visual question answering for UAVs.
ScholarMind
Research Productivity Tool
An intelligent tool for improving research reading efficiency, including paper crawling, content analysis, report generation, and research idea summarization.
Honors and Awards
- Second Place Award (2/62), RoCo Challenge @ AAAI 2026: Robotic Collaborative Assembling for Human-Centered Manufacturing, Team Leader
- Grand Prize, Beihang "Feng Ru Cup" Creativity Track
- Ranked 7/180 during undergraduate study
- First-Place Academic Scholarship, Beihang University
- First-Class Scholarship for Ph.D. Freshmen, Beihang University
Academic Service
- Reviewer for CVPR, ICCV, ECCV, ICLR, NeurIPS, ICML, and ACM MM