AI Can Learn Scientific Taste

TL;DR

This paper introduces a reinforcement learning framework enabling AI to develop scientific taste by learning from community feedback, leading to proposals with higher potential impact and surpassing state-of-the-art language models.

Contribution

It presents a novel training paradigm, RLCF, for teaching AI scientific taste through preference modeling and alignment, a previously underexplored area.

Findings

Scientific Judge outperforms SOTA LLMs in preference tasks.

Scientific Thinker proposes higher-impact research ideas.

Model generalizes across fields and future years.

Abstract

Great scientists have strong judgement and foresight, closely tied to what we call scientific taste. Here, we use the term to refer to the capacity to judge and propose research ideas with high potential impact. However, most relative research focuses on improving an AI scientist's executive capability, while enhancing an AI's scientific taste remains underexplored. In this work, we propose Reinforcement Learning from Community Feedback (RLCF), a training paradigm that uses large-scale community signals as supervision, and formulate scientific taste learning as a preference modeling and alignment problem. For preference modeling, we train Scientific Judge on 700K field- and time-matched pairs of high- vs. low-citation papers to judge ideas. For preference alignment, using Scientific Judge as a reward model, we train a policy model, Scientific Thinker, to propose research ideas with high potential impact. Experiments show Scientific Judge outperforms SOTA LLMs (e.g., GPT-5.2, Gemini 3 Pro) and generalizes to future-year test, unseen fields, and peer-review preference. Furthermore, Scientific Thinker proposes research ideas with higher potential impact than baselines. Our findings show that AI can learn scientific taste, marking a key step toward reaching human-level AI scientists.