Meta新模型帮AI研究代理智能选择实验，比传统方法快30%，省1/3算力
Meta提出AI研究偏好模型，解决研究代理缺乏原创性和实验选择问题。该模型基于冻结预训练LLM构建，推理模型评估候选计划、代码和已执行解决方案，代理模型额外运行小规模实验。在AIRS-Bench上，平均分从0.684提升至0.711和0.729，仅需15小时达到无引导代理24小时性能，使用不到三分之二执行预算。
Super interesting paper from Meta. Long-horizon research agents are coming. But one common problem...
Super interesting paper from Meta. Long-horizon research agents are coming. But one common problem with research agents today is the lack of originality and how to decide what experiments are worth exploring. An AI research agent can propose far more experiments than it can afford to run, so the problem is not idea generation, it's deciding which candidates get GPU time. AI Research Preference Models is trained to predict which candidate solution is most promising before any of them execute. Two variants, both built on frozen pretrained LLMs. An inference-only model reasons over candidate plans, code, and previously executed solutions. An agentic model additionally runs small-scale pilot experiments before committing budget. Dropped into the AIRA-dojo agent and measured on AIRS-Bench, average normalized score moves from 0.684 to 0.711 and 0.729. Both variants reach the unguided agent's 24-hour performance in roughly 15 hours, using less than two-thirds of its execution budget, and together set new state of the art on two AIRS-Bench tasks. Paper: arxiv.org/abs/2608.13940 Chat with Paper: academy.dair.ai/papers/ai-rese… 💬 4 🔄 2 ❤️ 15 👀 1153 📊 8 ⚡