@inproceedings{coca-etal-2025-pytod,
title = "{P}y{TOD}: Programmable Task-Oriented Dialogue with Execution Feedback",
author = "Coca, Alexandru and
Tseng, Bo-Hsiang and
Boothroyd, Peter and
Cheng, Jianpeng and
Zhang, Zhenxing and
Gaynor, Mark and
Stacey, Joe and
Guigue, Tristan and
Mart{\'i}nez Alonso, H{\'e}ctor and
{\'O} S{\'e}aghdha, Diarmuid and
Johannsen, Anders",
editor = "B{\'e}chet, Fr{\'e}d{\'e}ric and
Lef{\`e}vre, Fabrice and
Asher, Nicholas and
Kim, Seokhwan and
Merlin, Teva",
booktitle = "Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue",
month = aug,
year = "2025",
address = "Avignon, France",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.sigdial-1.6/",
pages = "93--112",
abstract = "Programmable task-oriented dialogue (TOD) agents enable language models to follow structured dialogue policies, but their effectiveness hinges on accurate dialogue state tracking (DST). We present PyTOD, an agent that generates executable code to track dialogue state and uses policy and execution feedback for efficient error correction. To achieve this, PyTOD employs a simple constrained decoding approach, using a language model instead of grammar rules to follow API schemata. This leads to state-of-the-art DST performance on the challenging SGD benchmark. Our experiments show that PyTOD surpasses strong baselines in both accuracy and cross-turn consistency, demonstrating the effectiveness of execution-aware state tracking."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="coca-etal-2025-pytod">
<titleInfo>
<title>PyTOD: Programmable Task-Oriented Dialogue with Execution Feedback</title>
</titleInfo>
<name type="personal">
<namePart type="given">Alexandru</namePart>
<namePart type="family">Coca</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bo-Hsiang</namePart>
<namePart type="family">Tseng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Peter</namePart>
<namePart type="family">Boothroyd</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jianpeng</namePart>
<namePart type="family">Cheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhenxing</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mark</namePart>
<namePart type="family">Gaynor</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Joe</namePart>
<namePart type="family">Stacey</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tristan</namePart>
<namePart type="family">Guigue</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Héctor</namePart>
<namePart type="family">Martínez Alonso</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Diarmuid</namePart>
<namePart type="family">Ó Séaghdha</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anders</namePart>
<namePart type="family">Johannsen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-08</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue</title>
</titleInfo>
<name type="personal">
<namePart type="given">Frédéric</namePart>
<namePart type="family">Béchet</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Fabrice</namePart>
<namePart type="family">Lefèvre</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nicholas</namePart>
<namePart type="family">Asher</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Seokhwan</namePart>
<namePart type="family">Kim</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Teva</namePart>
<namePart type="family">Merlin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Avignon, France</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Programmable task-oriented dialogue (TOD) agents enable language models to follow structured dialogue policies, but their effectiveness hinges on accurate dialogue state tracking (DST). We present PyTOD, an agent that generates executable code to track dialogue state and uses policy and execution feedback for efficient error correction. To achieve this, PyTOD employs a simple constrained decoding approach, using a language model instead of grammar rules to follow API schemata. This leads to state-of-the-art DST performance on the challenging SGD benchmark. Our experiments show that PyTOD surpasses strong baselines in both accuracy and cross-turn consistency, demonstrating the effectiveness of execution-aware state tracking.</abstract>
<identifier type="citekey">coca-etal-2025-pytod</identifier>
<location>
<url>https://aclanthology.org/2025.sigdial-1.6/</url>
</location>
<part>
<date>2025-08</date>
<extent unit="page">
<start>93</start>
<end>112</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T PyTOD: Programmable Task-Oriented Dialogue with Execution Feedback
%A Coca, Alexandru
%A Tseng, Bo-Hsiang
%A Boothroyd, Peter
%A Cheng, Jianpeng
%A Zhang, Zhenxing
%A Gaynor, Mark
%A Stacey, Joe
%A Guigue, Tristan
%A Martínez Alonso, Héctor
%A Ó Séaghdha, Diarmuid
%A Johannsen, Anders
%Y Béchet, Frédéric
%Y Lefèvre, Fabrice
%Y Asher, Nicholas
%Y Kim, Seokhwan
%Y Merlin, Teva
%S Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue
%D 2025
%8 August
%I Association for Computational Linguistics
%C Avignon, France
%F coca-etal-2025-pytod
%X Programmable task-oriented dialogue (TOD) agents enable language models to follow structured dialogue policies, but their effectiveness hinges on accurate dialogue state tracking (DST). We present PyTOD, an agent that generates executable code to track dialogue state and uses policy and execution feedback for efficient error correction. To achieve this, PyTOD employs a simple constrained decoding approach, using a language model instead of grammar rules to follow API schemata. This leads to state-of-the-art DST performance on the challenging SGD benchmark. Our experiments show that PyTOD surpasses strong baselines in both accuracy and cross-turn consistency, demonstrating the effectiveness of execution-aware state tracking.
%U https://aclanthology.org/2025.sigdial-1.6/
%P 93-112
Markdown (Informal)
[PyTOD: Programmable Task-Oriented Dialogue with Execution Feedback](https://aclanthology.org/2025.sigdial-1.6/) (Coca et al., SIGDIAL 2025)
ACL
- Alexandru Coca, Bo-Hsiang Tseng, Peter Boothroyd, Jianpeng Cheng, Zhenxing Zhang, Mark Gaynor, Joe Stacey, Tristan Guigue, Héctor Martínez Alonso, Diarmuid Ó Séaghdha, and Anders Johannsen. 2025. PyTOD: Programmable Task-Oriented Dialogue with Execution Feedback. In Proceedings of the 26th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 93–112, Avignon, France. Association for Computational Linguistics.