Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2602.23280 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866917547992940544 |
|---|---|
| author | Viswanath, Hrishikesh Lu, Juanwu Bukhari, S. Talha Chauhan, Mihir Conover, Damon Wang, Ziran Bera, Aniket |
| author_facet | Viswanath, Hrishikesh Lu, Juanwu Bukhari, S. Talha Chauhan, Mihir Conover, Damon Wang, Ziran Bera, Aniket |
| contents | Offline goal-conditioned reinforcement learning (GCRL) learns goal-reaching behaviors from static datasets, but accurate value estimation remains challenging under limited state-action coverage. Existing physics-informed approaches address this by imposing pointwise distance-like geometric constraints derived from Hamilton--Jacobi--Bellman (HJB) optimality principles, often through first-order partial differential equations such as the Eikonal equation. However, enforcing local consistency through explicit differential structure can become unstable in complex, high-dimensional environments. Our key insight is to instead reinterpret distance-like constraints as an expectation over a local spatial measure. By aggregating constraints over this measure rather than evaluating them pointwise, the objective acts as a spatial mollifier, inducing distance-like value geometry without requiring expensive differential operators. We refer to this as Mollified Value Learning (MVL). Experiments across navigation and high-dimensional robotic manipulation tasks show that MVL learns structured, value representations, improving goal-reaching performance, when used with implicit value representation learning methods. Open-source codes are available at https://github.com/HrishikeshVish/MVL. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_23280 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Mollified Value Learning Viswanath, Hrishikesh Lu, Juanwu Bukhari, S. Talha Chauhan, Mihir Conover, Damon Wang, Ziran Bera, Aniket Machine Learning Robotics Offline goal-conditioned reinforcement learning (GCRL) learns goal-reaching behaviors from static datasets, but accurate value estimation remains challenging under limited state-action coverage. Existing physics-informed approaches address this by imposing pointwise distance-like geometric constraints derived from Hamilton--Jacobi--Bellman (HJB) optimality principles, often through first-order partial differential equations such as the Eikonal equation. However, enforcing local consistency through explicit differential structure can become unstable in complex, high-dimensional environments. Our key insight is to instead reinterpret distance-like constraints as an expectation over a local spatial measure. By aggregating constraints over this measure rather than evaluating them pointwise, the objective acts as a spatial mollifier, inducing distance-like value geometry without requiring expensive differential operators. We refer to this as Mollified Value Learning (MVL). Experiments across navigation and high-dimensional robotic manipulation tasks show that MVL learns structured, value representations, improving goal-reaching performance, when used with implicit value representation learning methods. Open-source codes are available at https://github.com/HrishikeshVish/MVL. |
| title | Mollified Value Learning |
| topic | Machine Learning Robotics |
| url | https://arxiv.org/abs/2602.23280 |