Visual Navigation with Spatial Attention

TL;DR

This paper introduces a novel attention-based reinforcement learning approach for object goal visual navigation, combining semantic and spatial information to improve navigation accuracy and achieve state-of-the-art results.

Contribution

The main contribution is a new attention probability model that encodes both semantic and spatial information, enhancing the agent's navigation policy.

Findings

The attention model improves navigation performance.

Achieves state-of-the-art results on standard datasets.

Enhances the agent's ability to locate objects efficiently.

Abstract

This work focuses on object goal visual navigation, aiming at finding the location of an object from a given class, where in each step the agent is provided with an egocentric RGB image of the scene. We propose to learn the agent's policy using a reinforcement learning algorithm. Our key contribution is a novel attention probability model for visual navigation tasks. This attention encodes semantic information about observed objects, as well as spatial information about their place. This combination of the "what" and the "where" allows the agent to navigate toward the sought-after object effectively. The attention model is shown to improve the agent's policy and to achieve state-of-the-art results on commonly-used datasets.