Reflecting After Learning for Understanding
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.
Cite
@article{arxiv.1910.08243,
title = {Reflecting After Learning for Understanding},
author = {Lee Martie and Mohammad Arif Ul Alam and Gaoyuan Zhang and Ryan R. Anderson},
journal= {arXiv preprint arXiv:1910.08243},
year = {2020}
}
Comments
Presented at the Advances in Cognitive Systems conference (http://www.cogsys.org/conference/2019) and to be published in the Advances in Cognitive Systems journal (http://www.cogsys.org/journal)