用预训练模型预测实验方案优劣，节省大量计算资源。
AI Research Preference Models
- 基于冻结语言模型构建推理与试运行双版本偏好模型
- 在基准测试中将平均得分从0.684提升至0.729，提速近一半
- 适合需要高效实验筛选的自动化研究团队
AI研究代理（AIRA）现已能完成从实验设计到评估的全流程。然而，前沿任务进展受限于评估成本——耗时数天的GPU运算。当代理可提出的候选方案远超其可执行数量时，研究效率取决于其如何分配有限的执行预算。本文提出人工智能研究偏好模型（RPM），可在不运行全部方案的前提下预测最具潜力的候选解。RPM基于冻结的预训练语言模型，有两种变体：仅推理模型，通过分析候选计划、代码和历史结果进行判断；以及额外执行小规模试运行的代理型模型。将两种模型集成至AIRA-dojo研究代理，在机器学习研究基准AIRS-Bench上测试，平均归一化得分分别提升至0.711和0.729。两者均在约15小时内达到未引导代理24小时的性能，且使用不到其三分之二的执行预算，并在两个任务上取得新的最先进结果。
原文摘要 · Abstract (English)
AI research agents (AIRA) can now carry machine learning experiments from proposal through implementation and evaluation. Yet progress on frontier tasks is throttled by the cost of evaluations that can consume days of GPU time. When an agent can propose far more candidates than it can afford to run, progress depends on its research preference: how it allocates a fixed execution budget across many candidates. We introduce AI Research Preference Models (RPMs) that predict which candidate solution is most promising, without paying the cost of running them all. We build RPMs from frozen pretrained language models in two variants: an inference-only model that reasons over candidate plans, code, and previously executed solutions, and an agentic model that additionally runs small-scale pilot experiments. Integrated into the AIRA-dojo research agent and evaluated on the machine learning research benchmark AIRS-Bench, the two variants increase the average normalized score from 0.684 to 0.711 and 0.729, respectively. Both reach the unguided agent's 24-hour performance in roughly 15 hours, using less than two-thirds of its execution budget, and together yield new state-of-the-art results on two AIRS-Bench tasks.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。