Online Optimization of Curriculum Learning Schedules using Evolutionary Optimization
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866929456217587712 |
|---|---|
| author | Jiwatode, Mohit Schlecht, Leon Dockhorn, Alexander |
| author_facet | Jiwatode, Mohit Schlecht, Leon Dockhorn, Alexander |
| contents | We propose RHEA CL, which combines Curriculum Learning (CL) with Rolling Horizon Evolutionary Algorithms (RHEA) to automatically produce effective curricula during the training of a reinforcement learning agent. RHEA CL optimizes a population of curricula, using an evolutionary algorithm, and selects the best-performing curriculum as the starting point for the next training epoch. Performance evaluations are conducted after every curriculum step in all environments. We evaluate the algorithm on the \textit{DoorKey} and \textit{DynamicObstacles} environments within the Minigrid framework. It demonstrates adaptability and consistent improvement, particularly in the early stages, while reaching a stable performance later that is capable of outperforming other curriculum learners. In comparison to other curriculum schedules, RHEA CL has been shown to yield performance improvements for the final Reinforcement learning (RL) agent at the cost of additional evaluation during training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_06068 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Online Optimization of Curriculum Learning Schedules using Evolutionary Optimization Jiwatode, Mohit Schlecht, Leon Dockhorn, Alexander Artificial Intelligence Neural and Evolutionary Computing We propose RHEA CL, which combines Curriculum Learning (CL) with Rolling Horizon Evolutionary Algorithms (RHEA) to automatically produce effective curricula during the training of a reinforcement learning agent. RHEA CL optimizes a population of curricula, using an evolutionary algorithm, and selects the best-performing curriculum as the starting point for the next training epoch. Performance evaluations are conducted after every curriculum step in all environments. We evaluate the algorithm on the \textit{DoorKey} and \textit{DynamicObstacles} environments within the Minigrid framework. It demonstrates adaptability and consistent improvement, particularly in the early stages, while reaching a stable performance later that is capable of outperforming other curriculum learners. In comparison to other curriculum schedules, RHEA CL has been shown to yield performance improvements for the final Reinforcement learning (RL) agent at the cost of additional evaluation during training. |
| title | Online Optimization of Curriculum Learning Schedules using Evolutionary Optimization |
| topic | Artificial Intelligence Neural and Evolutionary Computing |
| url | https://arxiv.org/abs/2408.06068 |