Bregman divergence choice yields balanced vs. highly imbalanced 4-block partitions for Markov chains
michaelchchoi · x · 2026-10-22
An experiment by michaelchchoi on speeding up Markov chains via Bregman divergences: choosing different φ(u) generators produces very different "optimal" 4-block partitions of the same Gibbs target.
- φ(u)=u³ favors a balanced partition (25% mass per block)
- φ(u)=u^1.5 favors an imbalanced one, with one block holding 99.86% of the π-mass
Using lazy Metropolis as baseline, the results show divergence selection can systematically tune block structure for partitioned MCMC acceleration.
More from Research
- 36 preprints in 3 months: BootLoops AI harness drives cross-field science — DeryaTR_ · 2026-10-22
- Meta and MSU paper: LLM-built reasoning graphs are unreliable and their errors leak into confidence scores — rohanpaul_ai · 2026-10-22
- New paper: can first-price auctions efficiently allocate queries among strategic AI agents? — _onionesque · 2026-10-22
- RUC's LoopRec Brings Item-Level Adaptive Recursion with Early-Exit Gates to Sequential Recommendation — Zhanshuo Cao · 2026-10-22
- Anthropic's Agent4Sci: Frontier Agents Beat SOTA on 20 Scientific ML Benchmarks, but Only via Engineering Tricks — Anthropic · 2026-10-22
- NUS Paper Proposes PAPAV: A Unified Capability Framework from Computer-Use to Robot-Use Agents — showlab · 2026-10-22