Entropy hypocoercivity and query complexity for the bouncy particle sampler
We study convergence in relative entropy of the bouncy particle sampler with Gaussian velocity refreshment, for a target $e^{-U}$ on $\mathbb R^d$ with $U$ $m$-strongly convex and $L$-smooth, and condition number $κ=L/m$. For initial laws with bounded density relative to equilibrium, we show that the relative entropy decays exponentially up to an arbitrary additive error $δ$, with a rate that depends only logarithmically on the density bound and on $δ$. From a Gaussian cold start and with a suitably chosen refreshment rate, the relative entropy falls below $δ$ after time $\widetilde…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2610.25354
curl -X POST /api/v1/records/arxiv/2610.25354/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2610.25354/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'