analiza aiqbench
This research introduces CURL, a novel adapter that leverages frozen large language models to enhance the stability of heterogeneous treatment effect estimation by dynamically allocating semantic capacity based on estimator uncertainty. By separating representation pathways for assignment and heterogeneity tasks, the method improves performance across multiple benchmarks without requiring fine-tuning of the underlying LLM. The work represents a significant step forward in applying generative AI capabilities to rigorous causal inference problems like precision medicine and personalized marketing.
Uncertainty-Guided LLM Semantic Augmentation for Heterogeneous Treatment Effect Estimation
arXiv:2607.26599v1 Announce Type: new Abstract: Estimating heterogeneous treatment effects is central to targeted interventions, such as personalized promotions and precision medicine. We focus on the conditional average treatment effect (CATE), a standard estimand for characterizing such heterogeneity. Even under standard identification conditions, finite-sample CATE estimation requires learning the nuisance structure for covariate adjustment and treatment-effect heterogeneity, often together with an effective representation of X. Raw numerical and categorical encodings can leave semantic relations and higher-order interactions implicit, making this joint task locally unstable. A motivating study further shows that this instability appears through partially separable assignment- and heterogeneity-side channels. Building on this observation, we propose CURL (Causal Uncertainty-guided Representation Learning), a plug-in adapter that uses estimator uncertainty to allocate pretrained semantic capacity to locally unstable units. CURL queries a frozen LLM through two role-conditioned prompts, constructs assignment- and heterogeneity-oriented representations from the observed covariates, and routes them through separated pathways. On four benchmarks, CURL improves ten host learners in most settings, while ablation, refinement-dynamics, route-reassignment, and probe analyses support the intended design and roles of the two channels.
Abstract:Estimating heterogeneous treatment effects is central to targeted interventions, such as personalized promotions and precision medicine. We focus on the conditional average treatment effect (CATE), a standard estimand for characterizing such heterogeneity. Even under standard identification conditions, finite-sample CATE estimation requires learning the nuisance structure for covariate adjustment and treatment-effect heterogeneity, often together with an effective representation of X. Raw numerical and categorical encodings can leave semantic relations and higher-order interactions implicit, making this joint task locally unstable. A motivating study further shows that this instability appears through partially separable assignment- and heterogeneity-side channels. Building on this observation, we propose CURL (Causal Uncertainty-guided Representation Learning), a plug-in adapter that uses estimator uncertainty to allocate pretrained semantic capacity to locally unstable units. CURL queries a frozen LLM through two role-conditioned prompts, constructs assignment- and heterogeneity-oriented representations from the observed covariates, and routes them through separated pathways. On four benchmarks, CURL improves ten host learners in most settings, while ablation, refinement-dynamics, route-reassignment, and probe analyses support the intended design and roles of the two channels.
Machine Learning (cs.LG)
arXiv:2607.26599 [cs.LG]
(or arXiv:2607.26599v1 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2607.26599 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 267 cuvinte
Subiecte conexe
Optimizing Clinical Trial Protocols Using EHR-Derived Heterogeneous Treatment Effects
arXiv · 21 iul. 2026
UG-UMRE: Uncertainty-Guided Modality Augmentation and Distributional Calibration for Unified Multimodal Relation Extraction
arXiv · 6 aug. 2026
Beyond "What to Retrieve": Uncertainty in Retrieval-Augmented Code Generation
arXiv · 29 iul. 2026
C-MIG: Multi-view Information Gain-based Retrieval-Augmented Generation for Clinical Diagnosis Reasoning
arXiv · 5 aug. 2026