Posts by Collection

portfolio

publications

NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents

Published in ICLR 2026, 2026

We introduce NewtonBench, a benchmark of 324 scientific law discovery tasks across 12 physics domains. Counterfactual law shifts yield scalable, scientifically relevant, memorization-resistant problems, while interactive model discovery requires agents to experiment on simulated systems rather than fit static tables.

Recommended citation: Tianshi Zheng*, Kelvin Kiu-Wai Tam*, Newt Hue-Nam K. Nguyen*, et al. (2026). "NewtonBench: Benchmarking Generalizable Scientific Law Discovery in LLM Agents." ICLR. *Equal contribution.
Download Paper

SciResearcher: Scaling Deep Research Agents for Frontier Scientific Reasoning

Published in arXiv preprint, 2026

SciResearcher is a fully automated framework for constructing frontier-science training data from heterogeneous academic sources. SciResearcher-8B sets a new state of the art at the 8B scale on HLE-Bio/Chem-Gold and improves substantially on SuperGPQA-Hard-Biology and TRQA-Literature.

Recommended citation: Tianshi Zheng, Rui Wang, Xiyun Li, Kelvin Kiu-Wai Tam, Newt Hue-Nam K. Nguyen, Wei Fan, Yangqiu Song, Tianqing Fang (2026). "SciResearcher: Scaling Deep Research Agents for Frontier Scientific Reasoning." arXiv preprint arXiv:2605.01489.
Download Paper

talks

teaching

Teaching experience 1

Undergraduate course, University 1, Department, 2014

This is a description of a teaching experience. You can use markdown like any other post.

Teaching experience 2

Workshop, University 1, Department, 2015

This is a description of a teaching experience. You can use markdown like any other post.