{"id":"arxiv/2611.01436","url":"https://arxiv.org/abs/2611.01436","tags":["stat.ME","math.ST","stat.ML"],"score":0,"title":"Transport for Conditional Sampling: Foundations and Applications","domain":"stat.ME","source":"arxiv","authors":["Ricardo Baptista","Benjamin Bloem-Reddy","Bamdad Hosseini"],"reviews":0,"subject":"stat.ME","summary":"Conditional sampling is central to Bayesian inference, generative modeling, and causal inference: given a joint law for random variables $X$ and $Y$ and an observation $Y=y$, one seeks to sample from the conditional law of $X\\mid Y=y$. Even when the joint distribution is accurately represented, this task can be computationally difficult as it may require evaluating intractable normalizing constants or solving a new inference problem for each observation. Measure transport provides an approach to this problem by identifying a map with triangular structure that directly pushes forward a…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-11-01","reactions":{}}