Dianbo Liu is an Assistant Professor, National University of Singapore. He leads the Artificial Scientific Intelligence (ASI) Team at NUS to develop machine learning methods to enable AI to conduct autonomous research.
The ASI team focuses on the following three themes of research:
- Autoresearch: Enable AI to conduct autoresearch that improves AI itself.
- AI as science: Treat AI as a scientific subject to systematically understand the fundamentals of AI.
- AI as co-scientist: Develop AI to work with humans in biomedical science development.
Check more details about our team at https://www.asintelligence.xyz/
Selected publications:
- Chen, T., Lin, B., Yuan, Z., Zou, Q., He, H., Goyal, A., Ong, Y.-S., Liu, D. (2026). HypoSpace: Evaluating LLM Creativity as Set-Valued Hypothesis Generators under Underdetermination. ICML 2026 (spotlight)
- Sinha, A., Elango, S., Liu, D.(2026). Expected Return Causes Outcome-Level Mode Collapse in Reinforcement Learning and How to Fix It with Inverse Probability Scaling. ICML 2026
- (As co-corresponding co-last author )Xu, T., Peng, Q., Li, H., Wang, M., Zhang, J., Yu, M., Chong, C., Xue, C. C.,Zhang, X., He, H., Fu, H., Liu, D.*, Cheng, C.-Y.* (2026). Re-timing Mortality Risk through Individualized Modeling of Lifespan Trajectories. Nature Biomedical Engineering 2026
- Vetcha, N. and Liu, D.(2026) ‘SOLAR: A Self-Optimizing Open-Ended Autonomous Agent for Lifelong Learning and Continual Adaptation’, Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026). Main conference
- Dianbo Liu, Won-Yong Shin, Eli Sprecher, Kathleen Conroy, Omar Santiago, Gal Wachtel, Mauricio Santillana ”Machine learning approaches to predicting no-shows in pediatric medical appointment.” NPJ digital medicine 5.1 (2022): 1-11