Generalized Clustering by Learning to Optimize Expected Normalized Cuts

TL;DR

This paper presents a novel end-to-end deep learning approach for clustering that directly optimizes normalized cuts, achieving state-of-the-art results and strong generalization across diverse datasets without labeled data.

Contribution

It introduces a differentiable loss function for normalized cuts and a model that directly outputs cluster assignments, improving unsupervised clustering performance and generalization.

Findings

Achieves state-of-the-art results on MNIST, Reuters, CIFAR-10, and CIFAR-100.

Outperforms previous methods by up to 10.9% on benchmarks.

Demonstrates superior generalization ability, up to 21.9% better than recent approaches.

Abstract

We introduce a novel end-to-end approach for learning to cluster in the absence of labeled examples. Our clustering objective is based on optimizing normalized cuts, a criterion which measures both intra-cluster similarity as well as inter-cluster dissimilarity. We define a differentiable loss function equivalent to the expected normalized cuts. Unlike much of the work in unsupervised deep learning, our trained model directly outputs final cluster assignments, rather than embeddings that need further processing to be usable. Our approach generalizes to unseen datasets across a wide variety of domains, including text, and image. Specifically, we achieve state-of-the-art results on popular unsupervised clustering benchmarks (e.g., MNIST, Reuters, CIFAR-10, and CIFAR-100), outperforming the strongest baselines by up to 10.9%. Our generalization results are superior (by up to 21.9%) to the recent top-performing clustering approach with the ability to generalize.