Impact of Feedback Type on Explanatory Interactive Learning

TL;DR

This paper compares the effectiveness of two user feedback types in Explanatory Interactive Learning for image classification, finding that instructing models to ignore spurious features improves accuracy more than focusing on valid features.

Contribution

It provides a comparative analysis of feedback types in XIL, highlighting the superior impact of ignoring spurious features on model performance and explanation accuracy.

Findings

Ignoring spurious features leads to better classification accuracy.

Focusing on valid features results in lower explanation accuracy.

Feedback type significantly influences XIL effectiveness.

Abstract

Explanatory Interactive Learning (XIL) collects user feedback on visual model explanations to implement a Human-in-the-Loop (HITL) based interactive learning scenario. Different user feedback types will have different impacts on user experience and the cost associated with collecting feedback since different feedback types involve different levels of image annotation. Although XIL has been used to improve classification performance in multiple domains, the impact of different user feedback types on model performance and explanation accuracy is not well studied. To guide future XIL work we compare the effectiveness of two different user feedback types in image classification tasks: (1) instructing an algorithm to ignore certain spurious image features, and (2) instructing an algorithm to focus on certain valid image features. We use explanations from a Gradient-weighted Class Activation Mapping (GradCAM) based XIL model to support both feedback types. We show that identifying and annotating spurious image features that a model finds salient results in superior classification and explanation accuracy than user feedback that tells a model to focus on valid image features.