Low-shot Object Learning with Mutual Exclusivity Bias

TL;DR

This paper formalizes mutual exclusivity bias for low-shot object learning, introduces a new dataset, and develops a method that outperforms existing models in associating unknown objects with categories from minimal examples.

Contribution

It presents the first computational framework for mutual exclusivity bias, along with a dataset, baselines, and a novel method for low-shot object learning.

Findings

The proposed method outperforms state-of-the-art models in low-shot accuracy.

A new dataset and data generation pipeline for mutual exclusivity bias in object learning.

Comprehensive analysis of factors affecting low-shot learning difficulty.

Abstract

This paper introduces Low-shot Object Learning with Mutual Exclusivity Bias (LSME), the first computational framing of mutual exclusivity bias, a phenomenon commonly observed in infants during word learning. We provide a novel dataset, comprehensive baselines, and a state-of-the-art method to enable the ML community to tackle this challenging learning task. The goal of LSME is to analyze an RGB image of a scene containing multiple objects and correctly associate a previously-unknown object instance with a provided category label. This association is then used to perform low-shot learning to test category generalization. We provide a data generation pipeline for the LSME problem and conduct a thorough analysis of the factors that contribute to its difficulty. Additionally, we evaluate the performance of multiple baselines, including state-of-the-art foundation models. Finally, we present a baseline approach that outperforms state-of-the-art models in terms of low-shot accuracy.