[2603.27114] Maximin Learning of Individualized Treatment Effect on Multi-Domain Outcomes

Computer Science > Machine Learning arXiv:2603.27114 (cs) [Submitted on 28 Mar 2026] Title:Maximin Learning of Individualized Treatment Effect on Multi-Domain Outcomes Authors:Yuying Lu, Wenbo Fei, Yuanjia Wang, Molei Liu View a PDF of the paper titled Maximin Learning of Individualized Treatment Effect on Multi-Domain Outcomes, by Yuying Lu and 2 other authors View PDF HTML (experimental) Abstract:Precision mental health requires treatment decisions that account for heterogeneous symptoms reflecting multiple clinical domains. However, existing methods for estimating individualized treatment effects (ITE) rely on a single summary outcome or a specific set of observed symptoms or measures, which are sensitive to symptom selection and limit generalizability to unmeasured yet clinically relevant domains. We propose DRIFT, a new maximin framework for estimating robust ITEs from high-dimensional item-level data by leveraging latent factor representations and adversarial learning. DRIFT learns latent constructs via generalized factor analysis, then constructs an anchored on-target uncertainty set that extrapolates beyond the observed measures to approximate the broader hyper-population of potential outcomes. By optimizing worst-case performance over this uncertainty set, DRIFT yields ITEs that are robust to underrepresented or unmeasured domains. We further show that DRIFT is invariant to admissible reparameterizations of the latent factors and admits a closed-form maximin solut...