Unsupervised Class-Incremental Learning Through Confusion

TL;DR

This paper presents an unsupervised class-incremental learning method that detects novel classes by exploiting network confusion, significantly improving continual learning without labeled data across multiple image benchmarks.

Contribution

It introduces a novel unsupervised novelty detection technique based on network confusion, enhanced by class imbalance, for incremental learning without supervision.

Findings

Effective across multiple image classification benchmarks

Improves detection of novel classes without labels

Enhances continual learning performance

Abstract

While many works on Continual Learning have shown promising results for mitigating catastrophic forgetting, they have relied on supervised training. To successfully learn in a label-agnostic incremental setting, a model must distinguish between learned and novel classes to properly include samples for training. We introduce a novelty detection method that leverages network confusion caused by training incoming data as a new class. We found that incorporating a class-imbalance during this detection method substantially enhances performance. The effectiveness of our approach is demonstrated across a set of image classification benchmarks: MNIST, SVHN, CIFAR-10, CIFAR-100, and CRIB.