Search papers, labs, and topics across Lattice.
This paper establishes finite-time convergence rates of $\tilde{O}(1/\sqrt{n})$ for two model-free fixed-point algorithms designed to optimize discounted exponential utility in reinforcement learning, addressing the complexities introduced by its nonlinear structure. By employing parameter-free stepsizes and leveraging properties of the underlying power-law operator, the authors overcome significant challenges related to the update equations and tracking errors. These results represent the first finite-time guarantees for model-free reinforcement learning under discounted exponential utility, enhancing the understanding of risk-sensitive decision-making in sequential tasks.
Finite-time convergence rates for risk-sensitive reinforcement learning algorithms reveal that model-free approaches can achieve robust performance without complex parameter tuning.
Discounted exponential utility provides a principled criterion for risk-sensitive sequential decision-making, but its nonlinear structure complicates reinforcement learning. A recent work \citep{thoppe2026reinforcement} addressed this difficulty by introducing a Bellman-compatible surrogate and two model-free fixed-point algorithms for optimizing it over stationary policies. However, their main convergence results are asymptotic. In this work, we establish finite-time rates of $\tilde{O} (1/\sqrt{n})$ for the aforementioned two algorithms under asynchronous Markovian sampling, where $n$ is the iteration index and $\tilde{O}$ hides logarithmic expressions. Importantly, we employ parameter-free choices for the stepsize parameter to derive these rate results. For the algorithmically simpler one-timescale method, the main challenge is that its update equation is not directly aligned with the contraction geometry of its underlying power-law operator. We overcome this mismatch by exploiting the boundedness, monotonicity, and homogeneity of the operator to obtain a local pseudo-contraction property for the relative-error dynamics. We then use a Moreau-envelope-based Lyapunov function and Polyak--Ruppert averaging to obtain the stated convergence rate with parameter-free stepsizes. For the two-timescale method, the main challenge is to control a tracking error on the faster timescale. These results provide the first finite-time guarantees for model-free discounted exponential-utility reinforcement learning.
