GeoExplorer: Active Geo-localization with Curiosity-Driven Exploration
This addresses robustness issues in geo-localization for applications like robotics and navigation, though it is incremental as it builds on existing reinforcement learning frameworks.
The paper tackles the problem of active geo-localization where current methods struggle with robustness and generalization when distance estimation is difficult or in unseen scenarios, and proposes GeoExplorer, which uses curiosity-driven exploration to achieve effective localization across diverse benchmarks.
Active Geo-localization (AGL) is the task of localizing a goal, represented in various modalities (e.g., aerial images, ground-level images, or text), within a predefined search area. Current methods approach AGL as a goal-reaching reinforcement learning (RL) problem with a distance-based reward. They localize the goal by implicitly learning to minimize the relative distance from it. However, when distance estimation becomes challenging or when encountering unseen targets and environments, the agent exhibits reduced robustness and generalization ability due to the less reliable exploration strategy learned during training. In this paper, we propose GeoExplorer, an AGL agent that incorporates curiosity-driven exploration through intrinsic rewards. Unlike distance-based rewards, our curiosity-driven reward is goal-agnostic, enabling robust, diverse, and contextually relevant exploration based on effective environment modeling. These capabilities have been proven through extensive experiments across four AGL benchmarks, demonstrating the effectiveness and generalization ability of GeoExplorer in diverse settings, particularly in localizing unfamiliar targets and environments.