Expected flow networks in stochastic environments and two-player zero-sum games
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2023
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913264286302208 |
|---|---|
| author | Jiralerspong, Marco Sun, Bilun Vucetic, Danilo Zhang, Tianyu Bengio, Yoshua Gidel, Gauthier Malkin, Nikolay |
| author_facet | Jiralerspong, Marco Sun, Bilun Vucetic, Danilo Zhang, Tianyu Bengio, Yoshua Gidel, Gauthier Malkin, Nikolay |
| contents | Generative flow networks (GFlowNets) are sequential sampling models trained to match a given distribution. GFlowNets have been successfully applied to various structured object generation tasks, sampling a diverse set of high-reward objects quickly. We propose expected flow networks (EFlowNets), which extend GFlowNets to stochastic environments. We show that EFlowNets outperform other GFlowNet formulations in stochastic tasks such as protein design. We then extend the concept of EFlowNets to adversarial environments, proposing adversarial flow networks (AFlowNets) for two-player zero-sum games. We show that AFlowNets learn to find above 80% of optimal moves in Connect-4 via self-play and outperform AlphaZero in tournaments. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2310_02779 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Expected flow networks in stochastic environments and two-player zero-sum games Jiralerspong, Marco Sun, Bilun Vucetic, Danilo Zhang, Tianyu Bengio, Yoshua Gidel, Gauthier Malkin, Nikolay Machine Learning Computer Science and Game Theory Generative flow networks (GFlowNets) are sequential sampling models trained to match a given distribution. GFlowNets have been successfully applied to various structured object generation tasks, sampling a diverse set of high-reward objects quickly. We propose expected flow networks (EFlowNets), which extend GFlowNets to stochastic environments. We show that EFlowNets outperform other GFlowNet formulations in stochastic tasks such as protein design. We then extend the concept of EFlowNets to adversarial environments, proposing adversarial flow networks (AFlowNets) for two-player zero-sum games. We show that AFlowNets learn to find above 80% of optimal moves in Connect-4 via self-play and outperform AlphaZero in tournaments. |
| title | Expected flow networks in stochastic environments and two-player zero-sum games |
| topic | Machine Learning Computer Science and Game Theory |
| url | https://arxiv.org/abs/2310.02779 |