Continual Prototype Evolution: Learning Online from Non-Stationary Data Streams

TL;DR

This paper introduces a novel online continual learning system that evolves class prototypes in a shared latent space, effectively handling non-stationary data streams and imbalanced data with state-of-the-art results.

Contribution

It presents a system for online prototype evolution in a shared latent space, incorporating a new objective function and a learner-evaluator framework for data stream learning.

Findings

Achieves state-of-the-art performance on eight benchmarks.

Effectively handles highly imbalanced data streams.

Reduces catastrophic forgetting in non-stationary environments.

Abstract

Attaining prototypical features to represent class distributions is well established in representation learning. However, learning prototypes online from streaming data proves a challenging endeavor as they rapidly become outdated, caused by an ever-changing parameter space during the learning process. Additionally, continual learning does not assume the data stream to be stationary, typically resulting in catastrophic forgetting of previous knowledge. As a first, we introduce a system addressing both problems, where prototypes evolve continually in a shared latent space, enabling learning and prediction at any point in time. In contrast to the major body of work in continual learning, data streams are processed in an online fashion, without additional task-information, and an efficient memory scheme provides robustness to imbalanced data streams. Besides nearest neighbor based prediction, learning is facilitated by a novel objective function, encouraging cluster density about the class prototype and increased inter-class variance. Furthermore, the latent space quality is elevated by pseudo-prototypes in each batch, constituted by replay of exemplars from memory. As an additional contribution, we generalize the existing paradigms in continual learning to incorporate data incremental learning from data streams by formalizing a two-agent learner-evaluator framework. We obtain state-of-the-art performance by a significant margin on eight benchmarks, including three highly imbalanced data streams.