Streaming Inference for Infinite Non-Stationary Clustering

Rylan Schaeffer, Gabrielle Kaili-May Liu, Yilun Du, Scott Linderman, Ila Rani Fiete

Conference on Lifelong Learning Agents Accepted

August 2022

Bayesian Nonparametrics Streaming Inference Clustering Online Learning

PDF

Abstract

We develop streaming inference algorithms for infinite non-stationary clustering, enabling online learning of cluster structure that can change over time without requiring fixed cluster numbers.

Summary

Streaming inference algorithms for infinite non-stationary clustering - handling evolving cluster structures online.

The Problem

Traditional clustering algorithms assume a fixed number of clusters and stationary data. In real-world scenarios, data often arrives in streams and cluster structure can evolve over time.

Our Approach

We develop streaming inference algorithms for infinite non-stationary clustering, which:

• Handle online data streams
• Automatically determine the number of clusters
• Adapt to changing cluster structure over time

This enables lifelong learning systems that can continuously update their understanding of data structure.

---

See the full research page for more details.