Meta-learning representations for clustering with infinite Gaussian mixture models

TL;DR

This paper introduces a meta-learning approach that trains neural networks to produce representations optimized for clustering with infinite Gaussian mixture models, improving clustering performance on unseen data.

Contribution

It presents a novel meta-learning framework that directly optimizes representations for clustering via a differentiable approximation of the ARI and VB inference, enabling better generalization.

Findings

Higher adjusted Rand index than existing methods

Effective on both text and image datasets

Meta-learned representations improve clustering performance

Abstract

For better clustering performance, appropriate representations are critical. Although many neural network-based metric learning methods have been proposed, they do not directly train neural networks to improve clustering performance. We propose a meta-learning method that train neural networks for obtaining representations such that clustering performance improves when the representations are clustered by the variational Bayesian (VB) inference with an infinite Gaussian mixture model. The proposed method can cluster unseen unlabeled data using knowledge meta-learned with labeled data that are different from the unlabeled data. For the objective function, we propose a continuous approximation of the adjusted Rand index (ARI), by which we can evaluate the clustering performance from soft clustering assignments. Since the approximated ARI and the VB inference procedure are differentiable, we can backpropagate the objective function through the VB inference procedure to train the neural networks. With experiments using text and image data sets, we demonstrate that our proposed method has a higher adjusted Rand index than existing methods do.