Preprint

A Human-Augmenting Agentic Workflow for Observational Causal Inference

Winston Chou Adrien Alexandre Lars Olds Yi Zhang Nathan Kallus

Jul 2026 · 0 citations · 21 references

Mathematics

TL;DR

Across numerous evaluations, `oci-agent`, an open-source Python package that implements a human-in-the-loop agentic workflow for observational causal inference, outperforms less structured baselines while remaining competitive with hand-tuned benchmarks.

Abstract

Data analysis agents are becoming increasingly common tools for applied and scientific research. Yet, for highly specialized tasks such as Observational Causal Inference (OCI), human oversight remains necessary to ensure the validity of results. We introduce `oci-agent`, an open-source Python package that implements a human-in-the-loop agentic workflow for observational causal inference. `oci-agent` is designed to automate vital but laborious aspects of applied causal inference, such as covariate balance checking, propensity score trimming, and sensitivity analysis, so that humans can focus on more nuanced tasks, such as framing questions, scrutinizing assumptions, and evaluating diagnostics and results. We initially open-sourced `oci-agent` in June 2026 with support for doubly robust learning of the average treatment effect of a single binary treatment. Since then, we have added support for heterogeneous treatment effect estimation and for multiple continuous treatments via partially linear models. In this paper, we describe the principles behind `oci-agent` and offer internal Netflix case studies and evaluations on public data of its capabilities. Across numerous evaluations, `oci-agent` outperforms less structured baselines while remaining competitive with hand-tuned benchmarks. `oci-agent` is used extensively for causal inference at Netflix and has orchestrated more than 100 analyses per month since its release in June.

arXiv Semantic Scholar

View source

Similar papers

#artificial intelligence Preprint Sep 2026

When and Why LLM Causal Priors Help: Closed-Loop Prior Selection for Amortized Causal Inference

A closed-loop prior selection framework is proposed that casts prior injection as a budget-constrained optimization over a candidate prior pool and shows that the use of LLM causal priors stops being manual trial and error and becomes an empirically verifiable selection problem.

Hao-Hao Zhou · 0 citations

#artificial intelligence Preprint Sep 2026

TRACE: Training Reasoning Agents for Causal Exploration with Synthesized Rewards

Results provide evidence that access to a scalable, objective training signal can be a more important constraint than model scale alone, and can make otherwise ambiguous diagnostic reasoning tasks amenable to scalable reinforcement learning.

Rui Sun, Zhan Shi, B. He · 0 citations

#machine learning Preprint Sep 2026

Causal Foundation Models

Causal foundation models are pretrained neural networks that estimate causal quantities, such as the average treatment effect, on entirely new datasets using in-context learning without requiring model updates.

Christopher Stith, Hossein Rahmani, Jesse C. Cresswell · 0 citations

Aug 2026

Self-supervised Causal Effects Estimation

Self-supervised Causal Effects Estimation is proposed, a novel framework that integrates causal priors with self-supervised learning to construct balanced and predictive representations for causal effects estimation that consistently outperforms state-of-the-art methods.

Xin-Shu Li, Shiyi Yang, Venus Haghighi et al. · ACM Transactions on Intellig... · 0 citations

Preprint Aug 2026

Interpretable Causal Discovery via Causal-Effect Constraints

This work considers the task of conditional causal discovery as a Bayesian inference problem, in which the posterior is targeted over causal graphs and parameters conditional on an event such as a causal-effect constraint, and adapts rare-event estimation techniques to perform inference the joint graph-parameter space.

Cixuan Zhang, Guy Van den Broeck, Benjie Wang · 0 citations

Open access Sep 2026

Causal Machine Learning for Discovering Actionable Insights in Observational Data

A unified causal machine learning framework that integrates structural causal modeling with data-driven estimation techniques to enable robust causal discovery and effect estimation and achieves higher causal discovery accuracy with improved precision and recall of causal edges.

Maria Ulfa, Marwan Alshar'e, Dharmesh Dhabliya et al. · Journal of Data Science · 0 citations