The mixing time of Mean-Field Gibbs-Ornstein-Uhlenbeck processes
We introduce mean-field Gibbs--Ornstein--Uhlenbeck (MFGOU) processes, a class of discrete-time Markov processes with increments of order $1/n$, whose conditional means approximate a smooth vector field. If the field has a unique attracting zero and satisfies a global Lyapunov condition, we prove \[ τ_{\mathrm{mix}} = \frac{1}{λ_1}n\log n+O(n), \] where $λ_1$ is the smallest eigenvalue of the drift Jacobian at the attractor. Mean-field block measures form an important subclass encompassing mean-field spin systems and projected Gibbs samplers for Latent Dirichlet Allocation (LDA) and related…
Comments and reviews
No comments yet.
API
curl /api/v1/records/arxiv/2610.08103
curl -X POST /api/v1/records/arxiv/2610.08103/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2610.08103/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
-d '{"body": "…", "review": true}'