Growing Representation Learning

TL;DR

This paper introduces GMAT, an attention-based Gaussian Mixture model that learns interpretable data representations and detects new classes without prior assumptions, enhancing continual learning and avoiding catastrophic forgetting.

Contribution

The paper presents GMAT, a novel method combining Gaussian Mixtures with neural architecture search for unsupervised and supervised continual learning with class detection.

Findings

GMAT effectively detects new classes without prior distribution assumptions.

The method can learn interpretable representations with or without labels.

GMAT avoids catastrophic forgetting through replay mechanisms.

Abstract

Machine learning continues to grow in popularity due to its ability to learn increasingly complex tasks. However, for many supervised models, the shift in a data distribution or the appearance of a new event can result in a severe decrease in model performance. Retraining a model from scratch with updated data can be resource intensive or impossible depending on the constraints placed on an organization or system. Continual learning methods attempt to adapt models to new classes instead of retraining. However, many of these methods do not have a detection method for new classes or make assumptions about the distribution of classes. In this paper, we develop an attention based Gaussian Mixture, called GMAT, that learns interpretable representations of data with or without labels. We incorporate this method with existing Neural Architecture Search techniques to develop an algorithm for detection new events for an optimal number of representations through an iterative process of training a growing. We show that our method is capable learning new representations of data without labels or assumptions about the distributions of labels. We additionally develop a method that allows our model to utilize labels to more accurately develop representations. Lastly, we show that our method can avoid catastrophic forgetting by replaying samples from learned representations.