Multi-task Curriculum Learning In A Complex, Visual, Hard-exploration Domain: Minecraft

Abstract

An important challenge in reinforcement learning is training agents that can solve a wide variety of tasks. If tasks depend on each other (e.g. needing to learn to walk before learning to run), curriculum learning can speed up learning by focusing on the next best task to learn. We explore curriculum learning in a complex, visual domain with many hard exploration challenges: Minecraft. We find that learning progress (defined as a change in success probability of a task) is a reliable measure of learnability for automatically constructing an effective curriculum. We introduce a learning-progress based curriculum and test it on a complex reinforcement learning problem (called "Simon Says") where an agent is instructed to obtain a desired goal item. Many of the required skills depend on each other. Experiments demonstrate that: (1) a within-episode exploration bonus for obtaining new items improves performance, (2) dynamically adjusting this bonus across training such that it only applies

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keykanitscheider2021multi

Related papers

• Curriculum Learning With A Progression Function (2020)0.00
• Improving Deep Reinforcement Learning In Minecraft With Action Advice (2019)9.03
• Learning Progress Driven Multi-agent Curriculum (2022)0.00
• Learning Curriculum Policies For Reinforcement Learning (2018)5.24
• Curriculum Learning For Cooperation In Multi-agent Reinforcement Learning (2023)0.00
• Exploratory Gradient Boosting For Reinforcement Learning In Complex Domains (2016)0.00
• Proximal Curriculum With Task Correlations For Deep Reinforcement Learning (2024)0.00
• Open-world Multi-task Control Through Goal-aware Representation Learning And Adaptive Horizon Prediction (2023)8.09