Stein Variational Distributional Reinforcement Learning: Score-Guided Particles for Multivariate Return Distributions

arXiv:2610.17925v1 Announce Type: new Abstract: Decisions that weigh several reward signals against each other require the joint distribution of the return vector. Categorical critics scale exponentially in the reward dimension, and quantiles have no canonical multivariate analogue. Multivariate distributional critics therefore represent each return distribution by M equal-weight particles trained with a discrepancy loss. Beyond their atoms, the particles provide no sampler and cannot resolve events rarer than 1/M. We propose Stein variational distributional RL (SVDRL). Its critic outputs Stein particles, which Stein variational gradient descent transports along a learned score field of the distributional Bellman target. We learn this field by bootstrapped denoising score matching across noise levels and tie it across the Bellman recursion by a consistency loss, which we prove is a non-expansive temporal-difference scheme in score space. The same field certifies the particles. Its reverse-diffusion sampler satisfies a supremal Wasserstein policy-evaluation bound that holds for bounded, possibly multimodal returns. The Stein particles inherit this bound up to a gap that we measure without ground truth, and the sampler answers tail-risk queries below the particle resolution at any sample size. Empirically, SVDRL is more accurate on rare events than particle critics with the same number of particles, and on par on MuJoCo continuous control.

Abstract:Decisions that weigh several reward signals against each other require the joint distribution of the return vector. Categorical critics scale exponentially in the reward dimension, and quantiles have no canonical multivariate analogue. Multivariate distributional critics therefore represent each return distribution by M equal-weight particles trained with a discrepancy loss. Beyond their atoms, the particles provide no sampler and cannot resolve events rarer than 1/M. We propose Stein variational distributional RL (SVDRL). Its critic outputs Stein particles, which Stein variational gradient descent transports along a learned score field of the distributional Bellman target. We learn this field by bootstrapped denoising score matching across noise levels and tie it across the Bellman recursion by a consistency loss, which we prove is a non-expansive temporal-difference scheme in score space. The same field certifies the particles. Its reverse-diffusion sampler satisfies a supremal Wasserstein policy-evaluation bound that holds for bounded, possibly multimodal returns. The Stein particles inherit this bound up to a gap that we measure without ground truth, and the sampler answers tail-risk queries below the particle resolution at any sample size. Empirically, SVDRL is more accurate on rare events than particle critics with the same number of particles, and on par on MuJoCo continuous control.

Machine Learning (cs.LG)

arXiv:2610.17925 [cs.LG]

(or arXiv:2610.17925v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2610.17925 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.

analiza aiqbench

Stein Variational Distributional Reinforcement Learning: Score-Guided Particles for Multivariate Return Distributions

Subiecte conexe