Learning Transferable Reward for Query Object Localization with Policy Adaptation

TL;DR

This paper introduces a reinforcement learning method for query object localization that uses a transferable reward signal, enabling effective policy adaptation to new environments and classes without extensive retraining.

Contribution

The paper presents a novel transferable reward formulation using ordinal metric learning, allowing test-time policy adaptation and class transfer in object localization tasks.

Findings

Outperforms fine-tuning on various datasets

Enables class transfer without retraining

Effective in corrupted and diverse datasets

Abstract

We propose a reinforcement learning based approach to query object localization, for which an agent is trained to localize objects of interest specified by a small exemplary set. We learn a transferable reward signal formulated using the exemplary set by ordinal metric learning. Our proposed method enables test-time policy adaptation to new environments where the reward signals are not readily available, and outperforms fine-tuning approaches that are limited to annotated images. In addition, the transferable reward allows repurposing the trained agent from one specific class to another class. Experiments on corrupted MNIST, CU-Birds, and COCO datasets demonstrate the effectiveness of our approach.