Online Unsupervised Learning of Visual Representations and Categories

TL;DR

This paper introduces an unsupervised online learning model that learns visual representations and categories from raw, nonstationary data streams without labels, outperforming existing self-supervised methods in category recognition.

Contribution

It proposes a prototype-based memory network with an online mixture model that forms new categories from single examples, handling imbalanced data and nonstationary environments.

Findings

Learned representations outperform state-of-the-art self-supervised methods in category recognition.

Model effectively forms new categories from single examples in an online setting.

Handles nonstationary data streams with imbalanced class distributions.

Abstract

Real world learning scenarios involve a nonstationary distribution of classes with sequential dependencies among the samples, in contrast to the standard machine learning formulation of drawing samples independently from a fixed, typically uniform distribution. Furthermore, real world interactions demand learning on-the-fly from few or no class labels. In this work, we propose an unsupervised model that simultaneously performs online visual representation learning and few-shot learning of new categories without relying on any class labels. Our model is a prototype-based memory network with a control component that determines when to form a new class prototype. We formulate it as an online mixture model, where components are created with only a single new example, and assignments do not have to be balanced, which permits an approximation to natural imbalanced distributions from uncurated raw data. Learning includes a contrastive loss that encourages different views of the same image to be assigned to the same prototype. The result is a mechanism that forms categorical representations of objects in nonstationary environments. Experiments show that our method can learn from an online stream of visual input data and its learned representations are significantly better at category recognition compared to state-of-the-art self-supervised learning methods.