PONI: Potential Functions for ObjectGoal Navigation with Interaction-free Learning
This addresses the problem of efficient training for robotic navigation tasks, offering a significant reduction in computational resources while maintaining performance.
The paper tackles the high computational cost of reinforcement learning for ObjectGoal navigation by proposing PONI, a modular approach that separates object search and navigation, achieving state-of-the-art results on Gibson and Matterport3D datasets with up to 1,600x less training cost.
State-of-the-art approaches to ObjectGoal navigation rely on reinforcement learning and typically require significant computational resources and time for learning. We propose Potential functions for ObjectGoal Navigation with Interaction-free learning (PONI), a modular approach that disentangles the skills of `where to look?' for an object and `how to navigate to (x, y)?'. Our key insight is that `where to look?' can be treated purely as a perception problem, and learned without environment interactions. To address this, we propose a network that predicts two complementary potential functions conditioned on a semantic map and uses them to decide where to look for an unseen object. We train the potential function network using supervised learning on a passive dataset of top-down semantic maps, and integrate it into a modular framework to perform ObjectGoal navigation. Experiments on Gibson and Matterport3D demonstrate that our method achieves the state-of-the-art for ObjectGoal navigation while incurring up to 1,600x less computational cost for training. Code and pre-trained models are available: https://vision.cs.utexas.edu/projects/poni/