Teaching Categories to Human Learners with Visual Explanations

TL;DR

This paper introduces a novel computer-assisted teaching framework that uses visual explanations to enhance human learning of image categories, demonstrating improved performance over traditional methods.

Contribution

It proposes a new teaching approach that provides interpretable visual explanations to human learners, improving concept acquisition beyond conventional instance-level feedback.

Findings

Learners achieve higher test accuracy with visual explanations.

Automatically generated explanations highlight key image parts responsible for labels.

The approach outperforms existing teaching methods in experiments.

Abstract

We study the problem of computer-assisted teaching with explanations. Conventional approaches for machine teaching typically only provide feedback at the instance level e.g., the category or label of the instance. However, it is intuitive that clear explanations from a knowledgeable teacher can significantly improve a student's ability to learn a new concept. To address these existing limitations, we propose a teaching framework that provides interpretable explanations as feedback and models how the learner incorporates this additional information. In the case of images, we show that we can automatically generate explanations that highlight the parts of the image that are responsible for the class label. Experiments on human learners illustrate that, on average, participants achieve better test set performance on challenging categorization tasks when taught with our interpretable approach compared to existing methods.