RLIE: Rule Generation with Logistic Regression, Iterative Refinement, and Evaluation for Large Language Models

Neural-symbolic Completed method paper

Yang Yang*, Hua XU*, Zhangyi Hu*, Yutao Yue (* equal contribution)

43rd International Conference on Machine Learning (ICML) 2026 Accepted

Summary This work constructs explainable rules for scientific discovery using LLMs. Specifically, we develop a unified framework integrating LLMs with probabilistic modeling through four stages: rule generation via LLM, weight learning through logistic regression, iterative refinement, and evaluation, achieving higher rule quality and better rule combination effects.

My role Equal-contribution core contributor; worked on probabilistic integration for combining LLM-generated rules, evaluation design, and paper writing.

[arXiv]

A Very Big Video Reasoning Suite

Cognitive Modeling Benchmark contribution

Maijunxian Wang, Ruisi Wang, Juyi Lin, ..., Hua XU, ..., Hokin Deng†

43rd International Conference on Machine Learning (ICML) 2026 Accepted

Summary A GrowAI project introducing the Very Big Video Reasoning (VBVR) suite, a large-scale dataset and benchmark for evaluating video models' spatiotemporal reasoning, generalization, and verifiable task-solving abilities.

My role GrowAI contributor and large-team co-author; helped frame cognitive-evaluation tasks for testing video generative models' reasoning abilities.

[Project] [arXiv]

Building Interpretable, Trustworthy Systems for Neural Signal Decoding

Cognitive Modeling Single-author research proposal

Hua XU

AAAI 2026 Undergraduate Consortium 2026 Accepted proposal

Summary Single-author Undergraduate Consortium research proposal on developing interpretable and trustworthy neural signal decoding systems.

My role Single-author undergraduate consortium proposal; formulated the problem framing and research agenda for interpretable, trustworthy neural signal decoding.