RANa: Retrieval-Augmented Navigation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866911080885780480 |
|---|---|
| author | Monaci, Gianluca Rezende, Rafael S. Deffayet, Romain Csurka, Gabriela Bono, Guillaume Déjean, Hervé Clinchant, Stéphane Wolf, Christian |
| author_facet | Monaci, Gianluca Rezende, Rafael S. Deffayet, Romain Csurka, Gabriela Bono, Guillaume Déjean, Hervé Clinchant, Stéphane Wolf, Christian |
| contents | Methods for navigation based on large-scale learning typically treat each episode as a new problem, where the agent is spawned with a clean memory in an unknown environment. While these generalization capabilities to an unknown environment are extremely important, we claim that, in a realistic setting, an agent should have the capacity of exploiting information collected during earlier robot operations. We address this by introducing a new retrieval-augmented agent, trained with RL, capable of querying a database collected from previous episodes in the same environment and learning how to integrate this additional context information. We introduce a unique agent architecture for the general navigation task, evaluated on ImageNav, Instance-ImageNav and ObjectNav. Our retrieval and context encoding methods are data-driven and employ vision foundation models (FM) for both semantic and geometric understanding. We propose new benchmarks for these settings and we show that retrieval allows zero-shot transfer across tasks and environments while significantly improving performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_03524 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | RANa: Retrieval-Augmented Navigation Monaci, Gianluca Rezende, Rafael S. Deffayet, Romain Csurka, Gabriela Bono, Guillaume Déjean, Hervé Clinchant, Stéphane Wolf, Christian Computer Vision and Pattern Recognition Information Retrieval Robotics Methods for navigation based on large-scale learning typically treat each episode as a new problem, where the agent is spawned with a clean memory in an unknown environment. While these generalization capabilities to an unknown environment are extremely important, we claim that, in a realistic setting, an agent should have the capacity of exploiting information collected during earlier robot operations. We address this by introducing a new retrieval-augmented agent, trained with RL, capable of querying a database collected from previous episodes in the same environment and learning how to integrate this additional context information. We introduce a unique agent architecture for the general navigation task, evaluated on ImageNav, Instance-ImageNav and ObjectNav. Our retrieval and context encoding methods are data-driven and employ vision foundation models (FM) for both semantic and geometric understanding. We propose new benchmarks for these settings and we show that retrieval allows zero-shot transfer across tasks and environments while significantly improving performance. |
| title | RANa: Retrieval-Augmented Navigation |
| topic | Computer Vision and Pattern Recognition Information Retrieval Robotics |
| url | https://arxiv.org/abs/2504.03524 |