AI Research Preference Models

Engineering notes

Engineering notes will be added by the aipentium editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为大语言模型、生成式AI、ChatGPT相关技术、计算机视觉、深度学习等高价值论文补充中文说明。

Original abstract

AI research agents (AIRA) can now propose, implement, and evaluate their own machine learning experiments, but progress on frontier tasks is throttled by cost: a candidate solution can be written in minutes, whereas evaluating it can take hours to days of GPU time. An agent can therefore propose far more candidates than it can afford to run, and its 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 of multiple candidate solutions are most worth executing, without paying the cost of executing them all. We build RPMs from frozen pretrained language models (with no task-specific training), in two forms: an inference-only model that reasons over candidate plans, code, and prior executed solutions, and an agentic model that additionally runs small-scale pilot experiments before deciding. We integrate both into the AIRA-dojo search agent and evaluate on AIRS-Bench, a recent benchmark of machine learning research tasks for AI research agents. The two variants raise the average normalized score from 0.684 to 0.711 and 0.729 respectively, and reach the unguided agent's 24-hour performance in roughly 15 hours, using less than two-thirds of its execution budget. Our best RPMs also yield new state-of-the-art results on two AIRS-Bench tasks.

Links and sources

• Official / arXiv page
• PDF from original source

Comments