@inproceedings{roy-etal-2025-causal,
title = "Causal-{LLM}: A Unified One-Shot Framework for Prompt- and Data-Driven Causal Graph Discovery",
author = "Roy, Amartya and
Devharish, N and
Ganguly, Shreya and
Ghosh, Kripabandhu",
editor = "Christodoulopoulos, Christos and
Chakraborty, Tanmoy and
Rose, Carolyn and
Peng, Violet",
booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
month = nov,
year = "2025",
address = "Suzhou, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.findings-emnlp.439/",
doi = "10.18653/v1/2025.findings-emnlp.439",
pages = "8259--8279",
ISBN = "979-8-89176-335-7",
abstract = "Current causal discovery methods using Large Language Models (LLMs) often rely on pairwise or iterative strategies, which fail to capture global dependencies, amplify local biases, and reduce overall accuracy. This work introduces a unified framework for one-step full causal graph discovery through: (1) \textbf{Prompt-based discovery} with in-context learning when node metadata is available, and (2) \textbf{Causal{\_}llm}, a data-driven method for settings without metadata. Empirical results demonstrate that the prompt-based approach outperforms state-of-the-art models (GranDAG, GES, ICA-LiNGAM) by approximately 40{\%} in edge accuracy on datasets like Asia and Sachs, while maintaining strong performance on more complex graphs (ALARM, HEPAR2). Causal{\_}llm consistently excels across all benchmarks, achieving 50{\%} faster inference than reinforcement learning-based methods and improving precision by 25{\%} in fairness-sensitive domains such as legal decision-making. We also introduce two domain-specific DAGs{---}one for bias propagation and another for legal reasoning under the Bhartiya Nyaya Sanhita{---}demonstrating LLMs' capability for systemic, real-world causal discovery."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="roy-etal-2025-causal">
<titleInfo>
<title>Causal-LLM: A Unified One-Shot Framework for Prompt- and Data-Driven Causal Graph Discovery</title>
</titleInfo>
<name type="personal">
<namePart type="given">Amartya</namePart>
<namePart type="family">Roy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">N</namePart>
<namePart type="family">Devharish</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shreya</namePart>
<namePart type="family">Ganguly</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kripabandhu</namePart>
<namePart type="family">Ghosh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: EMNLP 2025</title>
</titleInfo>
<name type="personal">
<namePart type="given">Christos</namePart>
<namePart type="family">Christodoulopoulos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tanmoy</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Carolyn</namePart>
<namePart type="family">Rose</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Violet</namePart>
<namePart type="family">Peng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Suzhou, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-335-7</identifier>
</relatedItem>
<abstract>Current causal discovery methods using Large Language Models (LLMs) often rely on pairwise or iterative strategies, which fail to capture global dependencies, amplify local biases, and reduce overall accuracy. This work introduces a unified framework for one-step full causal graph discovery through: (1) Prompt-based discovery with in-context learning when node metadata is available, and (2) Causal_llm, a data-driven method for settings without metadata. Empirical results demonstrate that the prompt-based approach outperforms state-of-the-art models (GranDAG, GES, ICA-LiNGAM) by approximately 40% in edge accuracy on datasets like Asia and Sachs, while maintaining strong performance on more complex graphs (ALARM, HEPAR2). Causal_llm consistently excels across all benchmarks, achieving 50% faster inference than reinforcement learning-based methods and improving precision by 25% in fairness-sensitive domains such as legal decision-making. We also introduce two domain-specific DAGs—one for bias propagation and another for legal reasoning under the Bhartiya Nyaya Sanhita—demonstrating LLMs’ capability for systemic, real-world causal discovery.</abstract>
<identifier type="citekey">roy-etal-2025-causal</identifier>
<identifier type="doi">10.18653/v1/2025.findings-emnlp.439</identifier>
<location>
<url>https://aclanthology.org/2025.findings-emnlp.439/</url>
</location>
<part>
<date>2025-11</date>
<extent unit="page">
<start>8259</start>
<end>8279</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Causal-LLM: A Unified One-Shot Framework for Prompt- and Data-Driven Causal Graph Discovery
%A Roy, Amartya
%A Devharish, N.
%A Ganguly, Shreya
%A Ghosh, Kripabandhu
%Y Christodoulopoulos, Christos
%Y Chakraborty, Tanmoy
%Y Rose, Carolyn
%Y Peng, Violet
%S Findings of the Association for Computational Linguistics: EMNLP 2025
%D 2025
%8 November
%I Association for Computational Linguistics
%C Suzhou, China
%@ 979-8-89176-335-7
%F roy-etal-2025-causal
%X Current causal discovery methods using Large Language Models (LLMs) often rely on pairwise or iterative strategies, which fail to capture global dependencies, amplify local biases, and reduce overall accuracy. This work introduces a unified framework for one-step full causal graph discovery through: (1) Prompt-based discovery with in-context learning when node metadata is available, and (2) Causal_llm, a data-driven method for settings without metadata. Empirical results demonstrate that the prompt-based approach outperforms state-of-the-art models (GranDAG, GES, ICA-LiNGAM) by approximately 40% in edge accuracy on datasets like Asia and Sachs, while maintaining strong performance on more complex graphs (ALARM, HEPAR2). Causal_llm consistently excels across all benchmarks, achieving 50% faster inference than reinforcement learning-based methods and improving precision by 25% in fairness-sensitive domains such as legal decision-making. We also introduce two domain-specific DAGs—one for bias propagation and another for legal reasoning under the Bhartiya Nyaya Sanhita—demonstrating LLMs’ capability for systemic, real-world causal discovery.
%R 10.18653/v1/2025.findings-emnlp.439
%U https://aclanthology.org/2025.findings-emnlp.439/
%U https://doi.org/10.18653/v1/2025.findings-emnlp.439
%P 8259-8279
Markdown (Informal)
[Causal-LLM: A Unified One-Shot Framework for Prompt- and Data-Driven Causal Graph Discovery](https://aclanthology.org/2025.findings-emnlp.439/) (Roy et al., Findings 2025)
ACL