Sample Complexity of Multi-task Reinforcement Learning
This provides theoretical guarantees for multi-task reinforcement learning, addressing a gap in analysis for applications where tasks are sampled from an unknown distribution.
The paper tackles the problem of transferring knowledge across a sequence of reinforcement learning tasks to reduce sample complexity, proving that under certain assumptions, per-task sample complexity is significantly reduced compared to single-task algorithms, with no negative transfer in the worst case.
Transferring knowledge across a sequence of reinforcement-learning tasks is challenging, and has a number of important applications. Though there is encouraging empirical evidence that transfer can improve performance in subsequent reinforcement-learning tasks, there has been very little theoretical analysis. In this paper, we introduce a new multi-task algorithm for a sequence of reinforcement-learning tasks when each task is sampled independently from (an unknown) distribution over a finite set of Markov decision processes whose parameters are initially unknown. For this setting, we prove under certain assumptions that the per-task sample complexity of exploration is reduced significantly due to transfer compared to standard single-task algorithms. Our multi-task algorithm also has the desired characteristic that it is guaranteed not to exhibit negative transfer: in the worst case its per-task sample complexity is comparable to the corresponding single-task algorithm.