WORLD MODELS · GENERATIVE AI
Tao Hu 胡涛
Full Professor
Huazhong University of Science and Technology
World models for interactive, closed-loop simulation.
I am a Full Professor at HUST and lead VIBE Lab. My research centers on world models: learning geometric and semantic representations, modeling how environments evolve, and generating controllable future states.
I received my Ph.D. from the University of Amsterdam, advised by Cees G. M. Snoek and Pascal Mettes, and was a postdoctoral researcher in Björn Ommer’s lab at LMU Munich.
Research Keywords
- Memory in World Model
- Efficiency in World Model
- Multiplayer World Models
- Rolling-Based, Continuation-Based World Model
- Loop Transformer
- Tokenizer-Free
- Continuous Discrete Diffusion
JOIN VIBE LAB
2028 级研究生招生
招收保研/推免硕士研究生 4 人、博士研究生 2 人，欢迎与我联系！
LATEST UPDATES
News
| Jul 27, 2026 | Selected as an Excellent Reviewer for ECCV 2026. |
|---|---|
| May 27, 2026 | Diversifying Similar Subjects for Text-to-image Synthesis with Self-Cross Diffusion Guidance is accepted by IJCV, congrats to Weimin, Jieke, and Meng. |
| Feb 22, 2026 | Guiding Token-Sparse Diffusion Models is accepted by CVPR 2026, Congrats to Felix and the team~ |
| Jan 22, 2026 | Purrception: Variational Flow Matching for Vector-Quantized Image Generation is accepted by ICLR 2026. |
| Aug 01, 2025 | TREAD and ArtFlow are accepted by ICCV 2025, Congrats to the team~ |
PAPERS
Recent preprints
BACKGROUND
Experience
Appointments
- Apr 2026–present
- Full Professor · HUSTGenerative artificial intelligence research.
- Sep 2026–present
- Chief Technical Expert · Dongfeng Motor R&D InstitutePart-time · World models and interactive closed-loop simulation.
- Nov 2023–Nov 2025
- Postdoctoral Researcher · LMU MunichOmmer-Lab · Björn Ommer.
- Feb 2023–Nov 2024
- Visiting Researcher · University of AmsterdamVISLAB · Cees G. M. Snoek.
Earlier research internships at AWS Rekognition (2020), Megvii Face++ (2017–2018), and Xiaomi (2017).
ACADEMIC COMMUNITY
Service & recognition
Service & mentoring
Co-organizer of the CVPR 2025 eVLM and ECCV 2024 Audio-Visual Generative Learning workshops; lead organizer of the NxtAim workshop on the theory and practice of Stable Diffusion.
Area Chair for CVPR and ECCV workshops in 2024, and Session Chair for an AAAI 2025 computer vision oral session.
Reviewer for CVPR, ICCV, ECCV, NeurIPS, ICLR, ICML, AAAI, and journals including T-PAMI and TMLR.
Teaching and co-supervision at LMU Munich and the University of Amsterdam, spanning computer vision, generative models, and representation learning.
Selected recognition
- ICML 2026 · Silver Reviewer Award
- NeurIPS 2024 · Top Reviewer
- CVPR 2023 · Doctoral Consortium travel support; selected among the top 13% of 411 applicants
