Saved in:
Bibliographic Details
Main Authors: Erwin, Kyle, Axelrod, Guy, Chang, Maria, Fokoue, Achille, Crouse, Maxwell, Dan, Soham, Gao, Tian, Uceda-Sosa, Rosario, Makondo, Ndivhuwo, Khan, Naweed, Gray, Alexander
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2501.11335
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912194581495808
author Erwin, Kyle
Axelrod, Guy
Chang, Maria
Fokoue, Achille
Crouse, Maxwell
Dan, Soham
Gao, Tian
Uceda-Sosa, Rosario
Makondo, Ndivhuwo
Khan, Naweed
Gray, Alexander
author_facet Erwin, Kyle
Axelrod, Guy
Chang, Maria
Fokoue, Achille
Crouse, Maxwell
Dan, Soham
Gao, Tian
Uceda-Sosa, Rosario
Makondo, Ndivhuwo
Khan, Naweed
Gray, Alexander
contents The task of policy compliance detection (PCD) is to determine if a scenario is in compliance with respect to a set of written policies. In a conversational setting, the results of PCD can indicate if clarifying questions must be asked to determine compliance status. Existing approaches usually claim to have reasoning capabilities that are latent or require a large amount of annotated data. In this work, we propose logical decomposition for policy compliance (LDPC): a neuro-symbolic framework to detect policy compliance using large language models (LLMs) in a few-shot setting. By selecting only a few exemplars alongside recently developed prompting techniques, we demonstrate that our approach soundly reasons about policy compliance conversations by extracting sub-questions to be answered, assigning truth values from contextual information, and explicitly producing a set of logic statements from the given policies. The formulation of explicit logic graphs can in turn help answer PCDrelated questions with increased transparency and explainability. We apply this approach to the popular PCD and conversational machine reading benchmark, ShARC, and show competitive performance with no task-specific finetuning. We also leverage the inherently interpretable architecture of LDPC to understand where errors occur, revealing ambiguities in the ShARC dataset and highlighting the challenges involved with reasoning for conversational question answering.
format Preprint
id arxiv_https___arxiv_org_abs_2501_11335
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Few-shot Policy (de)composition in Conversational Question Answering
Erwin, Kyle
Axelrod, Guy
Chang, Maria
Fokoue, Achille
Crouse, Maxwell
Dan, Soham
Gao, Tian
Uceda-Sosa, Rosario
Makondo, Ndivhuwo
Khan, Naweed
Gray, Alexander
Computation and Language
Artificial Intelligence
The task of policy compliance detection (PCD) is to determine if a scenario is in compliance with respect to a set of written policies. In a conversational setting, the results of PCD can indicate if clarifying questions must be asked to determine compliance status. Existing approaches usually claim to have reasoning capabilities that are latent or require a large amount of annotated data. In this work, we propose logical decomposition for policy compliance (LDPC): a neuro-symbolic framework to detect policy compliance using large language models (LLMs) in a few-shot setting. By selecting only a few exemplars alongside recently developed prompting techniques, we demonstrate that our approach soundly reasons about policy compliance conversations by extracting sub-questions to be answered, assigning truth values from contextual information, and explicitly producing a set of logic statements from the given policies. The formulation of explicit logic graphs can in turn help answer PCDrelated questions with increased transparency and explainability. We apply this approach to the popular PCD and conversational machine reading benchmark, ShARC, and show competitive performance with no task-specific finetuning. We also leverage the inherently interpretable architecture of LDPC to understand where errors occur, revealing ambiguities in the ShARC dataset and highlighting the challenges involved with reasoning for conversational question answering.
title Few-shot Policy (de)composition in Conversational Question Answering
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2501.11335