Open-world Multi-task Control Through Goal-aware Representation Learning And Adaptive Horizon Prediction

Abstract

We study the problem of learning goal-conditioned policies in Minecraft, a popular, widely accessible yet challenging open-ended environment for developing human-level multi-task agents. We first identify two main challenges of learning such policies: 1) the indistinguishability of tasks from the state distribution, due to the vast scene diversity, and 2) the non-stationary nature of environment dynamics caused by partial observability. To tackle the first challenge, we propose Goal-Sensitive Backbone (GSB) for the policy to encourage the emergence of goal-relevant visual state representations. To tackle the second challenge, the policy is further fueled by an adaptive horizon prediction module that helps alleviate the learning uncertainty brought by the non-stationary dynamics. Experiments on 20 Minecraft tasks show that our method significantly outperforms the best baseline so far; in many of them, we double the performance. Our ablation and exploratory studies then explain how our a

Stats

• citations11
• S2 citations—
• github stars0
• HF likes0
• heat score8.09
• arxiv keycai2023open

Related papers

• Multi-task Curriculum Learning In A Complex, Visual, Hard-exploration Domain: Minecraft (2021)0.00
• Backward Learning For Goal-conditioned Policies (2023)0.00
• Juewu-mc: Playing Minecraft With Sample-efficient Hierarchical Reinforcement Learning (2021)0.00
• Self-supervised Goal-reaching Results In Multi-agent Cooperation And Exploration (2025)0.00
• Learning, Fast And Slow: A Goal-directed Memory-based Approach For Dynamic Environments (2023)0.00
• Deep Decentralized Multi-task Multi-agent Reinforcement Learning Under Partial Observability (2017)0.00
• Improving Deep Reinforcement Learning In Minecraft With Action Advice (2019)9.03
• Understanding And Controlling A Maze-solving Policy Network (2023)0.00