Learning Novel Objects Continually Through Curiosity

TL;DR

This paper presents a curiosity-driven continual learning method for robots that improves object recognition by actively querying uncertain objects, outperforming traditional sampling methods on a benchmark dataset.

Contribution

It introduces a self-supervised technique for identifying uncertain objects using cluster representations, enhancing continual learning in robotic environments.

Findings

Outperforms random sampling in classification accuracy

Achieves better results than softmax-based uncertainty sampling

Learns more classes effectively in a continual learning setting

Abstract

Children learn continually by asking questions about the concepts they are most curious about. With robots becoming an integral part of our society, they must also learn unknown concepts continually by asking humans questions. The paper analyzes a recent state-of-the-art approach for continual learning. The paper further develops a self-supervised technique to find most of the uncertain objects in an environment by utilizing the cluster representation of the previously learned classes. We test our approach on a benchmark dataset for continual learning on robots. Our results show that our curiosity-driven continual learning approach beats random sampling and softmax-based uncertainty sampling in terms of classification accuracy and the total number of classes learned.