arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2301.05149cs.CLcs.AIcs.HCcs.LGcs.RO
收藏

定义、评估并改进指令生成模型的任务导向认知能力

Define, Evaluate, and Improve Task-Oriented Cognitive Capabilities for Instruction Generation Models

University of Maryland–College Park(马里兰大学帕克分校);Princeton University(普林斯顿大学);Microsoft Research(微软研究院)

Lingjun Zhao, Khanh Nguyen, Hal Daumé

AI总结:

本文提出任务导向认知能力概念并设计评估方案,发现导航指令生成模型语用能力严重不足,通过引入更好的听者模型使引导真实人类的成功率提升11%。

摘要:

近期研究通过为人类设计的心理测试来考察语言模型的认知能力。虽然这些研究有助于理解这些模型的通用能力,但无法保证一个具备足够能力通过这些测试的模型,在实际执行现实任务时真的会运用这些能力。在本工作中,我们形式化了任务导向认知能力,即语言模型在执行任务时所借助的类人认知能力。这些能力包括:(i)快速生成良好候选话语的能力(搜索能力);(ii)预测听者如何解读这些话语并选择最合适话语的能力(语用能力)。我们设计了一种评估方案,用于将语言模型的这些能力与人类的能力进行比较。将该方案应用于导航指令生成问题以考察多种模型后,我们发现它们的语用能力严重不足。这一发现促使我们为这些模型增加更好的听者模型,并在引导真实人类的成功率上获得了11%的显著提升。我们的工作主张采用一种有原则的流程来使语言模型与人类对齐,该流程包括:(i)形式化任务导向能力,(ii)设计量化其不足的方法,以及(iii)迭代改进这些能力。

英文摘要:

Recent work studies the cognitive capabilities of language models through psychological tests designed for humans. While these studies are helpful for understanding the general capabilities of these models, there is no guarantee that a model possessing sufficient capabilities to pass those tests would actually use those capabilities in performing real-life tasks. In this work, we formulate task-oriented cognitive capabilities, which are human-like cognitive capabilities that language models leverage to perform tasks. These capabilities are (i) the ability to quickly generate good candidate utterances (the search capability) (ii) the ability to predict how a listener interprets those utterances and choose the most appropriate one (the pragmatic capability). We design an evaluation scheme for comparing these capabilities of a language model with those of a human. Applying this scheme to examine various models in a navigation instruction generation problem, we find that their pragmatic capability is severely lacking. This insight leads us to augment them with better models of the listener and obtain a significant boost of 11% in success rate in guiding real humans. Our work advocates for having a principled procedure for aligning language models with humans that involves (i) formulating task-oriented capabilities, (ii) devising a method to quantify their deficiency, and (iii) iteratively improving them.

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。

↑