{"id":"arxiv/2610.01193","url":"https://arxiv.org/abs/2610.01193","tags":["cs.LG","cs.AI","math.ST","stat.ML"],"score":0,"title":"Counterfactual Generation via Flow Matching: Coupling-Sensitive End-to-End Rates","domain":"cs.LG","source":"arxiv","authors":["Yunrui Guan","Krishnakumar Balasubramanian","Shiva Prasad Kasiviswanathan"],"reviews":0,"subject":"cs.LG","summary":"Counterfactual generation seeks to sample outcomes under a hypothetical intervention or decision using observational data collected under the factual assignment mechanism. We develop a flow-matching approach that combines a sample-split, doubly robust training objective with a learned coupling between observed source outcomes and target outcomes drawn from a fitted conditional outcome model. To enable finite-step generation, we leverage a score-corrected stochastic sampler based on a Gaussian-smoothed interpolation. Our main theoretical contribution is a coupling-sensitive KL bound for…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-10-01","reactions":{}}