On Evaluating Abstraction and Analogy in Humans and Machines

IF 5.8 1区心理学 Q1 PSYCHOLOGY, MULTIDISCIPLINARY

Current Directions in Psychological Science Pub Date : 2026-04-22 DOI:10.1177/09637214261433528

Melanie Mitchell

引用次数: 0

Abstract

This article reviews two case studies in which AI systems were evaluated for abstraction and analogy-making capabilities and compared with those of humans. These studies illustrate how AI systems should be evaluated not only for accuracy on benchmark tasks but also for robustness to task variations and for insight into how the system is solving the tasks. These studies also illuminate the need for transparency, interpretability, and scientifically informed experimental methodology in AI evaluations.

查看原文本刊更多论文

人类与机器的抽象与类比评价

本文回顾了两个案例研究，其中评估了人工智能系统的抽象和类比能力，并将其与人类进行了比较。这些研究说明了如何评估人工智能系统，不仅要评估基准任务的准确性，还要评估任务变化的鲁棒性，并深入了解系统如何解决任务。这些研究还阐明了在人工智能评估中需要透明度、可解释性和科学的实验方法。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Current Directions in Psychological Science PSYCHOLOGY, MULTIDISCIPLINARY-

CiteScore

13.00

自引率

1.40%

发文量

期刊介绍： Current Directions in Psychological Science publishes reviews by leading experts covering all of scientific psychology and its applications. Each issue of Current Directions features a diverse mix of reports on various topics such as language, memory and cognition, development, the neural basis of behavior and emotions, various aspects of psychopathology, and theory of mind. These articles allow readers to stay apprised of important developments across subfields beyond their areas of expertise and bodies of research they might not otherwise be aware of. The articles in Current Directions are also written to be accessible to non-experts, making them ideally suited for use in the classroom as teaching supplements.