Sampling from multi-modal distributions with polynomial query complexity in fixed dimension via reverse diffusion

Adrien Vacher (ENSAE) · Omar Chehab (Carnegie Mellon University) · Anna Korba (GENES-CREST/ENSAE)
broad class of distributionsdiffusion processfixed dimensiongaussian mixturesintermediate score functionslog-smoothness assumptionmetastabilitymode locationsmulti-modal distributionspolynomial complexityquery complexitysampling algorithmscore functionsself-normalized monte carlotime-reversed diffusion

Even in low dimensions, sampling from multi-modal distributions is challenging. We provide the first sampling algorithm for a broad class of distributions