arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.02012cs.AIcs.ROcs.SE
收藏

预训练嵌入作为一种行为规约机制

Pretrained Embeddings as a Behavior Specification Mechanism

Carnegie Mellon University(卡内基梅隆大学);Scaled Foundations;University of Washington(华盛顿大学)

Parv Kapoor, Abigail Hammer, Ashish Kapoor, Karen Leung, Eunsuk Kang

AI总结:

提出引入嵌入作为规约语言一等构造,设计新型时序逻辑 ETL,通过理想与观测嵌入的距离形式化规约感知系统行为,并在基础模型驱动的机器人规划任务中验证其能有效引导系统表现出期望行为。

摘要:

我们提出了一种方法,用于形式化规约那些依赖感知模型与物理世界交互的系统的行为属性。核心思想是引入嵌入——真实世界概念的数学表示——作为规约语言中的一等构造,其中属性通过一对理想嵌入和观测嵌入之间的距离来表达。为实现此方法,我们提出了一种名为 Embedding Temporal Logic (ETL) 的新型时序逻辑,并描述了如何用它来表达 AI 系统比以往更广泛的属性。我们通过一项涉及由基础模型驱动的机器人规划任务的初步评估,展示了 ETL 的适用性;结果令人鼓舞,表明基于嵌入的规约可用于引导系统表现出期望行为。

英文摘要:

We propose an approach to formally specifying the behavioral properties of systems that rely on a perception model for interactions with the physical world. The key idea is to introduce embeddings -- mathematical representations of a real-world concept -- as a first-class construct in a specification language, where properties are expressed in terms of distances between a pair of ideal and observed embeddings. To realize this approach, we propose a new type of temporal logic called Embedding Temporal Logic (ETL), and describe how it can be used to express a wider range of properties about AI-enabled systems than previously possible. We demonstrate the applicability of ETL through a preliminary evaluation involving planning tasks in robots that are driven by foundation models; the results are promising, showing that embedding-based specifications can be used to steer a system towards desirable behaviors.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑