{"id":"arxiv/2608.00701","url":"https://arxiv.org/abs/2608.00701","tags":["stat.ME","cs.LG","math.ST"],"score":0,"title":"Augmented Inverse Hybrid Weighting: Robust Inference under Deterministic and Random Distribution Shifts","domain":"stat.ME","source":"arxiv","authors":["Ying Jin","Ying Jin","Dominik Rothenhäusler"],"reviews":0,"subject":"stat.ME","summary":"Reweighting source samples to match a target covariate distribution is a standard response to distribution shift when generalizing evidence from one population to another. This strategy is well suited to deterministic, learnable covariate discrepancies, but can be insufficient when source--target population differences also contain changes beyond covariate shift or when estimation of the density-ratio weights is unstable. To address this challenge, we introduce a new model that allows non-systematic changes between two population laws after systematic shifts are accounted for. Such residual…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-08-01","reactions":{}}