2021/02/17 by Nicola Milano, Milano, Nicola, Stefano Nolfi +1 · 1 citation
Computer Science · #Evolutionary Algorithms and Applications #FOS: Computer and information sciences #Metaheuristic Optimization Algorithms Research #Neural and Evolutionary Computing (cs.NE) #Reinforcement Learning in Robotics
paper · pdf · doi:10.48550/arxiv.2102.08849
openalex publication_date 2021/02/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We demonstrate how an evolutionary algorithm can be extended with a\ncurriculum learning process that selects automatically the environmental\nconditions in which the evolving agents are evaluated. The environmental\nconditions are selected so to adjust the level of difficulty to the ability\nlevel of the current evolving agents and so to challenge the weaknesses of the\nevolving agents. The method does not require domain knowledge and does not\nintroduce additional hyperparameters. The results collected on two benchmark\nproblems, that require to solve a task in significantly varying environmental\nconditions, demonstrate that the method proposed outperforms conventional\nalgorithms and generates solutions that are robust to variations\n