Adaptive Multi-Horizon Reinforcement Learning
For reinforcement learning practitioners, this method offers a way to automatically adapt discount factors in non-stationary environments, reducing manual tuning effort.
The paper proposes a multi-horizon reinforcement learning approach that adaptively selects and combines temporal horizons, eliminating the need for manual discount-factor tuning. It demonstrates robust adaptation to changing reward structures in MiniGrid environments, including continual learning settings with three sequentially changing tasks.
Effective decision-making in complex and changing environments requires balancing short-term and long-term consequences. In reinforcement learning (RL), this trade-off is typically controlled through a fixed discount factor, which imposes a single exponentially discounted temporal horizon. However, biological agents exhibit flexible and adaptive temporal discounting, suggesting that effective planning requires multiple timescales. Here, we propose a multi-horizon approach that adaptively selects and combines temporal horizons, enabling robust adaptation to changes in reward structure without manual discount-factor tuning. This flexibility makes the method particularly suitable for continual learning scenarios involving task switches and varying environmental configurations. Empirically, we demonstrate that our approach identifies effective discount factors across a range of MiniGrid environments, including continual settings composed of three sequentially changing tasks. These results suggest that adaptive temporal discounting can improve parameter efficiency and enhance adaptability in both artificial and biologically inspired learning systems.