Leveraging Graph Structures and Large Language Models for End-to-End Synthetic Task-Oriented Dialogues

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Medjad, Maya, Imbert, Hugo, Yun, Bruno, Szymocha, Raphaël, Armetta, Frédéric
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915113225682944
author Medjad, Maya
Imbert, Hugo
Yun, Bruno
Szymocha, Raphaël
Armetta, Frédéric
author_facet Medjad, Maya
Imbert, Hugo
Yun, Bruno
Szymocha, Raphaël
Armetta, Frédéric
contents Training task-oriented dialogue systems is both costly and time-consuming, due to the need for high-quality datasets encompassing diverse intents. Traditional methods depend on extensive human annotation, while recent advancements leverage large language models (LLMs) to generate synthetic data. However, these approaches often require custom prompts or code, limiting accessibility for non-technical users. We introduce GraphTOD, an end-to-end framework that simplifies the generation of task-oriented dialogues. Users can create dialogues by specifying transition graphs in JSON format. Our evaluation demonstrates that GraphTOD generates high-quality dialogues across various domains, significantly lowering the cost and complexity of dataset creation.
format Preprint
id arxiv_https___arxiv_org_abs_2501_11977
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Leveraging Graph Structures and Large Language Models for End-to-End Synthetic Task-Oriented Dialogues
Medjad, Maya
Imbert, Hugo
Yun, Bruno
Szymocha, Raphaël
Armetta, Frédéric
Computation and Language
Artificial Intelligence
Training task-oriented dialogue systems is both costly and time-consuming, due to the need for high-quality datasets encompassing diverse intents. Traditional methods depend on extensive human annotation, while recent advancements leverage large language models (LLMs) to generate synthetic data. However, these approaches often require custom prompts or code, limiting accessibility for non-technical users. We introduce GraphTOD, an end-to-end framework that simplifies the generation of task-oriented dialogues. Users can create dialogues by specifying transition graphs in JSON format. Our evaluation demonstrates that GraphTOD generates high-quality dialogues across various domains, significantly lowering the cost and complexity of dataset creation.
title Leveraging Graph Structures and Large Language Models for End-to-End Synthetic Task-Oriented Dialogues
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2501.11977