Introducing the structural bases of typicality effects in deep learning

TL;DR

This paper introduces a novel computational model inspired by human semantic category representation, enabling deep learning models to better capture abstract concepts like typicality and family resemblance.

Contribution

The paper proposes a new Computational Prototype Model (CPM) that enhances deep neural networks' ability to represent semantic categories with typicality and abstract meanings.

Findings

CPM improves semantic representation in image classification tasks.

The approach demonstrates effectiveness on datasets like ImageNet and Coco.

Results suggest enhanced abstraction capabilities in deep learning models.

Abstract

In this paper, we hypothesize that the effects of the degree of typicality in natural semantic categories can be generated based on the structure of artificial categories learned with deep learning models. Motivated by the human approach to representing natural semantic categories and based on the Prototype Theory foundations, we propose a novel Computational Prototype Model (CPM) to represent the internal structure of semantic categories. Unlike other prototype learning approaches, our mathematical framework proposes a first approach to provide deep neural networks with the ability to model abstract semantic concepts such as category central semantic meaning, typicality degree of an object's image, and family resemblance relationship. We proposed several methodologies based on the typicality's concept to evaluate our CPM-model in image semantic processing tasks such as image classification, a global semantic description, and transfer learning. Our experiments on different image datasets, such as ImageNet and Coco, showed that our approach might be an admissible proposition in the effort to endow machines with greater power of abstraction for the semantic representation of objects' categories.