Reflecting After Learning for Understanding

TL;DR

This paper introduces a novel meta-reasoning framework that unifies multiple image classification predictions into meaningful abstractions, improving interpretability and understanding of neural network outputs on ImageNet and live video data.

Contribution

It presents a new framework that uses meta-operations to unify and explain neural network predictions, enhancing interpretability in image classification systems.

Findings

Unifies 41% to 46% of predictions on ImageNet.

Achieves 67% to 75% unification when explanations are available.

Unifies 51% of predictions in live video scenarios.

Abstract

Today, image classification is a common way for systems to process visual content. Although neural network approaches to classification have seen great progress in reducing error rates, it is not clear what this means for a cognitive system that needs to make sense of the multiple and competing predictions from its own classifiers. As a step to address this, we present a novel framework that uses meta-reasoning and meta-operations to unify predictions into abstractions, properties, or relationships. Using the framework on images from ImageNet, we demonstrate systems that unify 41% to 46% of predictions in general and unify 67% to 75% of predictions when the systems can explain their conceptual differences. We also demonstrate a system in "the wild" by feeding live video images through it and show it unifying 51% of predictions in general and 69% of predictions when their differences can be explained conceptually by the system. In a survey given to 24 participants, we found that 87% of the unified predictions describe their corresponding images.