Unsupervised Online Learning for Robotic Interestingness with Visual Memory

TL;DR

This paper introduces an unsupervised online learning system with a novel visual memory architecture that enables robots to adaptively identify interesting scenes in real-time, outperforming existing methods in subterranean environments.

Contribution

It presents a novel translation-invariant visual memory and a three-stage architecture for long-term, short-term, and online learning in robotic interestingness detection.

Findings

20% higher accuracy than state-of-the-art unsupervised methods

Comparable performance to supervised methods in exploration tasks

Effective online adaptation in subterranean environments

Abstract

Autonomous robots frequently need to detect "interesting" scenes to decide on further exploration, or to decide which data to share for cooperation. These scenarios often require fast deployment with little or no training data. Prior work considers "interestingness" based on data from the same distribution. Instead, we propose to develop a method that automatically adapts online to the environment to report interesting scenes quickly. To address this problem, we develop a novel translation-invariant visual memory and design a three-stage architecture for long-term, short-term, and online learning, which enables the system to learn human-like experience, environmental knowledge, and online adaption, respectively. With this system, we achieve an average of 20% higher accuracy than the state-of-the-art unsupervised methods in a subterranean tunnel environment. We show comparable performance to supervised methods for robot exploration scenarios showing the efficacy of our approach. We expect that the presented method will play an important role in the robotic interestingness recognition exploration tasks.