Distributionally Robust Inverse Reinforcement Learning for Identifying Multi-Agent Coordinated Sensing
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866912040314994688 |
|---|---|
| author | Snow, Luke Krishnamurthy, Vikram |
| author_facet | Snow, Luke Krishnamurthy, Vikram |
| contents | We derive a minimax distributionally robust inverse reinforcement learning (IRL) algorithm to reconstruct the utility functions of a multi-agent sensing system. Specifically, we construct utility estimators which minimize the worst-case prediction error over a Wasserstein ambiguity set centered at noisy signal observations. We prove the equivalence between this robust estimation and a semi-infinite optimization reformulation, and we propose a consistent algorithm to compute solutions. We illustrate the efficacy of this robust IRL scheme in numerical studies to reconstruct the utility functions of a cognitive radar network from observed tracking signals. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2409_14542 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Distributionally Robust Inverse Reinforcement Learning for Identifying Multi-Agent Coordinated Sensing Snow, Luke Krishnamurthy, Vikram Machine Learning Multiagent Systems Signal Processing We derive a minimax distributionally robust inverse reinforcement learning (IRL) algorithm to reconstruct the utility functions of a multi-agent sensing system. Specifically, we construct utility estimators which minimize the worst-case prediction error over a Wasserstein ambiguity set centered at noisy signal observations. We prove the equivalence between this robust estimation and a semi-infinite optimization reformulation, and we propose a consistent algorithm to compute solutions. We illustrate the efficacy of this robust IRL scheme in numerical studies to reconstruct the utility functions of a cognitive radar network from observed tracking signals. |
| title | Distributionally Robust Inverse Reinforcement Learning for Identifying Multi-Agent Coordinated Sensing |
| topic | Machine Learning Multiagent Systems Signal Processing |
| url | https://arxiv.org/abs/2409.14542 |