Graphical Normalizing Flow

Graphical Normalizing Flows (GNFs) are a class of deep learning models used for causal inference, leveraging the flexibility of neural networks to represent complex probability distributions while incorporating causal relationships represented by directed acyclic graphs (DAGs). Current research focuses on applying GNFs to various problems, including offline reinforcement learning, sensitivity analysis under unobserved confounding, and counterfactual analysis for personalized policy interventions. This approach offers advantages over traditional methods by enabling flexible, semi-parametric estimation of causal effects, including individual-level predictions, and handling complex relationships without strong functional form assumptions, thereby improving the accuracy and applicability of causal inference across diverse fields.

Papers

November 2, 2024

Marginal Causal Flows for Validation and Inference
Daniel de Vassimon Manela, Laura Battaglia, Robin J. Evans
Scientific Inference Causal Discovery Causal Effect External Validation Causal View Unmeasured Confounding Graphical Normalizing Flow

May 6, 2024

Out-of-Distribution Adaptation in Offline RL: Counterfactual Reasoning via Causal Normalizing Flows
Minjae Cho, Jonathan P. How, Chuangchuang Sun
Offline Reinforcement Learning Policy Learning Counterfactual Reasoning Offline Policy Evaluation Causal Reinforcement Learning Graphical Normalizing Flow Graph Out of Distribution Adaptation

January 12, 2024

Deep Learning With DAGs
Sourabh Balgi, Adel Daoud, Jose M. Peña, Geoffrey T. Wodtke, Jesse Zhou
Deep Learning Causal Inference Causal Effect Causal Relationship Directed Acyclic Graph Parametric Causal Graphical Normalizing Flow

September 15, 2022

$\rho$-GNF : A Novel Sensitivity Analysis Approach Under Unobserved Confounders
Sourabh Balgi, Jose M. Peña, Adel Daoud
Causal Inference Causal Effect Unobserved Confounders Sensitivity Analysis Average Causal Effect Gaussian Copula Graphical Normalizing Flow

February 17, 2022

Counterfactual Analysis of the Impact of the IMF Program on Child Poverty in the Global-South Region using Causal-Graphical Normalizing Flows
Sourabh Balgi, Jose M. Peña, Adel Daoud
Causal Inference Counterfactual Inference Counterfactual Analysis Poverty Level Graphical Normalizing Flow Central Bank

February 7, 2022

Personalized Public Policy Analysis in Social Sciences using Causal-Graphical Normalizing Flows
Sourabh Balgi, Jose M. Pena, Adel Daoud
Causal Effect Social Science Counterfactual Inference Average Causal Effect Online Policy Customization Graphical Normalizing Flow