Yong Dai
AI researcher whose work has consistently centered on model pre-training and post-training, spanning language models, multimodal foundation models, agents, and embodied intelligence. I conducted research at Microsoft and Tencent AI Lab, where I worked on the training and adaptation of foundation models.
Current focus: embodied intelligence and world models — building the foundation brain models and interactive world models that let AI move from the digital world into the physical one.
Experience
- Tencent AI LabResearch Intern and Researcher
- Westlake UniversityVisiting Student
- Microsoft STCA NLPGResearch Intern
- Nuance collaborationProject Leader
News
| Date | Content |
|---|---|
| 2026 | 🎉 Pelican-Unify 1.0 released, a unified embodied intelligence model for understanding, reasoning, imagination and action |
| 2026 | 🎉 Papers accepted to ACL 2026, ICLR 2026, ICML 2026, EMNLP 2026, and ACM MM 2026 |
| Nov 2025 | 🎉 Pelican-VL 1.0 released, a foundation brain model for embodied intelligence |
| Sep 2025 | 🎉 Wow: a world omniscient world model through embodied interaction |
| Feb 2025 | 🎉 Two papers accepted to ACM MM 2025 |
| Sep 2024 | 🎉 Two papers accepted to NeurIPS 2024 |
| Sep 2024 | 🎉 One paper accepted to EMNLP 2024 Findings |
| Jun 2024 | 🎉 Three papers accepted to ACL 2024 |
| Jun 2023 | 🎉 SkillNet-X accepted to ICASSP 2024 |
| Dec 2022 | 🎉 Federated Learning + PLMs accepted to Findings of ACL 2023 |
| Oct 2022 | 🎉 Prompt-based Constrained Clustering accepted to Findings of EMNLP 2022 |
| Mar 2022 | 🎉 Whole Word Masking accepted to Findings of ACL 2022 |
| Mar 2022 | 🎉 Chinese GPT for Pinyin Input accepted to ACL 2022 |
| Oct 2020 | 🎉 Contextualize KBs with Transformer accepted to EMNLP 2021 (Oral) |
Site reach
- Visits
- —
- Countries
- —