Restricted Indian Buffet Processes

TL;DR

This paper introduces the Restricted Indian Buffet Process, a flexible nonparametric latent feature model that allows explicit control over the distribution of features per data point, enhancing modeling accuracy.

Contribution

It proposes the Restricted Indian Buffet Process, enabling arbitrary feature count distributions, and develops inference methods for this new model.

Findings

Allows explicit control over feature counts per data point

Develops MCMC and variational inference algorithms

Demonstrates improved modeling flexibility

Abstract

Latent feature models are a powerful tool for modeling data with globally-shared features. Nonparametric exchangeable models such as the Indian Buffet Process offer modeling flexibility by letting the number of latent features be unbounded. However, current models impose implicit distributions over the number of latent features per data point, and these implicit distributions may not match our knowledge about the data. In this paper, we demonstrate how the Restricted Indian Buffet Process circumvents this restriction, allowing arbitrary distributions over the number of features in an observation. We discuss several alternative constructions of the model and use the insights gained to develop Markov Chain Monte Carlo and variational methods for simulation and posterior inference.