TRESTLE: A Model of Concept Formation in Structured Domains
TRESTLE: A Model of Concept Formation in Structured Domains
复制标题
TRESTLE：结构化领域中概念形成的模型
DOI：
--
复制
发表时间：
2016
期刊：
影响因子：
--
通讯作者：
K. Koedinger
中科院分区：
文献类型：
--
作者：
Christopher Maclellan;Erik Harpstead;V. Aleven;K. Koedinger
关于概念形成的文献已经证明，人类能够以增量的方式学习概念，具有各种属性类型，并且在监督和非监督环境中都是如此。许多概念形成模型侧重于这些特征的子集，但没有一个模型解释了所有这些特征。在本文中，我们提出了Trestle，一个关于结构化领域中概率概念形成的增量描述，它统一了fiES的先验概念学习模型。Trestle的工作原理是创建一个分层分类树，该树可用于预测缺失的属性值，并将示例集聚集到概念上有意义的组中。它通过部分匹配新结构并将它们分类到其分类树中来更新其知识。最后，系统支持混合数据表示，包括名义、数字、关系和组件属性。我们评估了Trestle在有监督学习任务和无监督聚类任务上的性能。对于这两项任务，我们将其与非增量模型和人类参与者进行比较。我们fi发现，这种新的分类模型与非增量方法相比具有竞争力，并且更接近人类在这两个任务上的行为。这些结果是对Trestle能力的初步展示，并表明，通过考虑人类学习的关键特征，它可以比忽略它们的方法更好地模拟行为。
The literature on concept formation has demonstrated that humans are capable of learning concepts incrementally, with a variety of attribute types, and in both supervised and unsupervised settings. Many models of concept formation focus on a subset of these characteristics, but none account for all of them. In this paper, we present TRESTLE, an incremental account of probabilistic concept formation in structured domains that unifies prior concept learning models. TRESTLE works by creating a hierarchical categorization tree that can be used to predict missing attribute values and cluster sets of examples into conceptually meaningful groups. It updates its knowledge by partially matching novel structures and sorting them into its categorization tree. Finally, the system supports mixed-data representations, including nominal, numeric, relational, and component attributes. We evaluate TRESTLE’s performance on a supervised learning task and an unsupervised clustering task. For both tasks, we compare it to a nonincremental model and to human participants. We find that this new categorization model is competitive with the nonincremental approach and more closely approximates human behavior on both tasks. These results serve as an initial demonstration of TRESTLE’s capabilities and show that, by taking key characteristics of human learning into account, it can better model behavior than approaches that ignore them.