CLAMS: A System for Zero-Shot Model Selection for Clustering

TL;DR

This paper introduces CLAMS, an AutoML system that uses dataset similarity based on optimal transport to automatically select the best clustering algorithm, outperforming existing baselines.

Contribution

It presents a novel AutoML pipeline for clustering that leverages dataset similarity measures, expanding AutoML beyond supervised learning.

Findings

Outperforms multiple clustering baselines

Demonstrates effectiveness of similarity-based model selection

Establishes a new AutoML approach for clustering

Abstract

We propose an AutoML system that enables model selection on clustering problems by leveraging optimal transport-based dataset similarity. Our objective is to establish a comprehensive AutoML pipeline for clustering problems and provide recommendations for selecting the most suitable algorithms, thus opening up a new area of AutoML beyond the traditional supervised learning settings. We compare our results against multiple clustering baselines and find that it outperforms all of them, hence demonstrating the utility of similarity-based automated model selection for solving clustering applications.