arXiv cs.AI· Fernando Diaz·· 4 小时前AI 评分38
基于偏好的轨迹评估:用时间偏好替代二元成功率,将智能体评测平局率从约75%降至约35%
Offline Preference-Based Trajectory Evaluation
AI 导读
研究提出基于偏好的轨迹评估方法,通过比较轨迹间 time-to-return 曲线的时间偏好直接对比成对轨迹,替代将轨迹压缩为二元终端成功值的传统做法。在多种智能体与交互式基准上,成功率指标约75%的实例出现平局,而轨迹感知偏好将平局降至约35%,同时提升区分度、排名稳定性与数据效率。结果表明,常被归因于数据收集质量或问题难度的基准饱和,也可能源于评估指标本身的设计。
正文 · 原文
Abstract:Offline evaluation of agentic systems often collapses trajectories into a binary terminal success value, discarding information about partial progress. This, in turn, results in widespread ties that create statistical inefficiency by reducing effective sample size and weakening the ability to distinguish systems. We propose preference-based trajectory evaluation, an evaluation measure that compares pairs of trajectories directly by using temporal preferences between time-to-return profiles. We find that, across diverse agentic and interactive benchmarks, success-based metrics result in tied comparisons on roughly 75% of instances, while trajectory-aware preferences reduce ties to roughly 35%, improving discriminative power, ranking stability, and data efficiency. Our results suggest that benchmark saturation, often portrayed as the result of poor data collection or problem difficulty, may also be explained by the design of the evaluation measure.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2606.17541 [cs.LG] |
| (or arXiv:2606.17541v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2606.17541 arXiv-issued DOI via DataCite |
Submission history
From: Fernando Diaz [view email]
[v1]
Tue, 16 Jun 2026 05:42:19 UTC (449 KB)
[v2]
Mon, 16 Nov 2026 19:51:01 UTC (464 KB)
来源:arXiv cs.AI · arxiv.org