Suppr超能文献

鸽子(家鸽)和人类(智人)在基于范例和规则的分类任务中在局部和全局处理水平之间的转换。

Transfer between local and global processing levels by pigeons (Columba livia) and humans (Homo sapiens) in exemplar- and rule-based categorization tasks.

作者信息

Aust Ulrike, Braunöder Elisabeth

机构信息

Department of Cognitive Biology, University of Vienna.

出版信息

J Comp Psychol. 2015 Feb;129(1):1-16. doi: 10.1037/a0037691. Epub 2014 Aug 25.

Abstract

The present experiment investigated pigeons' and humans' processing styles-local or global-in an exemplar-based visual categorization task in which category membership of every stimulus had to be learned individually, and in a rule-based task in which category membership was defined by a perceptual rule. Group Intact was trained with the original pictures (providing both intact local and global information), Group Scrambled was trained with scrambled versions of the same pictures (impairing global information), and Group Blurred was trained with blurred versions (impairing local information). Subsequently, all subjects were tested for transfer to the 2 untrained presentation modes. Humans outperformed pigeons regarding learning speed and accuracy as well as transfer performance and showed good learning irrespective of group assignment, whereas the pigeons of Group Blurred needed longer to learn the training tasks than the pigeons of Groups Intact and Scrambled. Also, whereas humans generalized equally well to any novel presentation mode, pigeons' transfer from and to blurred stimuli was impaired. Both species showed faster learning and, for the most part, better transfer in the rule-based than in the exemplar-based task, but there was no evidence of the used processing mode depending on the type of task (exemplar- or rule-based). Whereas pigeons relied on local information throughout, humans did not show a preference for either processing level. Additional tests with grayscale versions of the training stimuli, with versions that were both blurred and scrambled, and with novel instances of the rule-based task confirmed and further extended these findings.

摘要

本实验在基于范例的视觉分类任务和基于规则的任务中,研究了鸽子和人类的加工方式——局部加工还是整体加工。在基于范例的视觉分类任务中,每个刺激的类别归属都必须单独学习;在基于规则的任务中,类别归属由一个感知规则定义。完整组用原始图片(提供完整的局部和整体信息)进行训练,打乱组用相同图片的打乱版本(破坏整体信息)进行训练,模糊组用模糊版本(破坏局部信息)进行训练。随后,所有受试者都接受测试,以考察向两种未训练呈现模式的迁移情况。在学习速度和准确性以及迁移表现方面,人类优于鸽子,并且无论分组如何,人类都表现出良好的学习能力,而模糊组的鸽子比完整组和打乱组的鸽子学习训练任务所需时间更长。此外,人类能够同样好地向任何新的呈现模式进行泛化,而鸽子在模糊刺激之间的迁移受到损害。在基于规则的任务中,两个物种都表现出更快的学习速度,并且在很大程度上迁移表现更好,但没有证据表明所使用的加工模式取决于任务类型(基于范例的或基于规则的)。鸽子始终依赖局部信息,而人类在两种加工水平上均未表现出偏好。对训练刺激的灰度版本、既模糊又打乱的版本以及基于规则任务的新实例进行的额外测试证实并进一步扩展了这些发现。

文献AI研究员

20分钟写一篇综述,助力文献阅读效率提升50倍。

立即体验

用中文搜PubMed

大模型驱动的PubMed中文搜索引擎

马上搜索

文档翻译

学术文献翻译模型,支持多种主流文档格式。

立即体验