GenAI-Powered Inference

2025/07/05 by Kosuke Imai, K. Nakamura, Kentaro Nakamura +2 · 1 voice · 1 citation
Biochemistry, Genetics and Molecular Biology · Computer Science · Mathematics · #AI in cancer detection #Anomaly Detection Techniques and Applications #Cell Image Analysis Techniques #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Methodology (stat.ME) #cs.LG #stat.ME #stat.ML

paper · pdf · doi:10.48550/arxiv.2507.03897

openalex publication_date 2025/07/05 · arxiv published 2025/07/05 · openalex created_date 2025/10/11 · openalex updated_date 2026/07/28 · arxiv created 2026/08/05 · arxiv updated 2026/08/06

Abstract

We introduce GenAI-Powered Inference (GPI), a statistical framework for both causal and predictive inference using unstructured data, including text and images. GPI leverages open-source Generative Artificial Intelligence (GenAI) models---such as large language models and diffusion models---not only to generate unstructured data at scale but also to extract low-dimensional representations that are guaranteed to capture their underlying structure. Applying machine learning to these representations, GPI enables estimation of causal effects while quantifying associated estimation uncertainty. Unlike existing approaches to representation learning, GPI does not require fine-tuning of generative models, making it computationally efficient and broadly accessible. We illustrate the versatility of the GPI framework through three applications: (1) estimating the effects of Chinese social media censorship while adjusting for textual confounders, (2) isolating the impact of specific image features from that of other correlated features in the same image, and (3) assessing the persuasiveness of political rhetoric. An open-source software package is available for implementing GPI.

Citations

• How Much Weak Overlap Can Doubly Robust T-Statistics Handle?
• Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments
• The Llama 3 Herd of Models
• Efficient adjustment for complex covariates: Gaining efficiency with DOPE
• Hyperparameter Tuning for Causal Inference with Double Machine Learning: A Simulation Study
• DoubleMLDeep: Estimation of Causal Effects with Multimodal Data
• Causal Estimation for Text Data with (Apparent) Overlap Violations
• Conceptualizing Treatment Leakage in Text-based Causal Inference
• High-Resolution Image Synthesis with Latent Diffusion Models
• High-Resolution Image Synthesis with Latent Diffusion Models
• Bayesian Topic Regression for Causal Inference
• Causal Effects of Linguistic Properties
• Optuna: A Next-generation Hyperparameter Optimization Framework
• Adapting Neural Networks for the Estimation of Treatment Effects
• Regularisation of Neural Networks by Enforcing Lipschitz Continuity
• How to Make Causal Inferences Using Texts
• Matching with Text Data: An Experimental Evaluation of Methods for Matching Documents and of Measuring Match Quality
• Overlap in Observational Studies with High-Dimensional Covariates
• Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learning
• Social Attributions from Faces: Determinants, Consequences, Accuracy, and Functional Significance
• How Censorship in China Allows Government Criticism but Silences Collective Expression
• RANK ANALYSIS OF INCOMPLETE BLOCK DESIGNS
• Causal Inference without Balance Checking: Coarsened Exact Matching

Cited by

• Causal Inferences from Digital Behavioral Data

Discussions

• GenAI-Powered Inference [hn, 1 points, 1 comments]

Related

• Causal Representation Learning with Generative Artificial Intelligence: Application to Texts as Treatments
• Tutorial on the Probabilistic Unification of Estimation Theory, Machine Learning, and Generative AI
• GenAI vs. Human Fact-Checkers: Accurate Ratings, Flawed Rationales
• Generative Modeling: A Review
• If generative AI is the answer, what is the question?