Unsupervised Dialogue Topic Segmentation with Topic-aware Utterance Representation

Haoyu Gao, Rui Wang, Ting-En Lin, Yuchuan Wu, Min Yang +2 more
5/4/2023
cs.CLcs.AI

Official code: AlibabaResearch/DAMO-ConvAI (1,602 stars, MIT, last commit 2026-01-22).

Abstract

Dialogue Topic Segmentation (DTS) plays an essential role in a variety of dialogue modeling tasks. Previous DTS methods either focus on semantic similarity or dialogue coherence to assess topic similarity for unsupervised dialogue segmentation. However, the topic similarity cannot be fully identified via semantic similarity or dialogue coherence. In addition, the unlabeled dialogue data, which contains useful clues of utterance relationships, remains underexploited. In this paper, we propose a novel unsupervised DTS framework, which learns topic-aware utterance representations from unlabeled dialogue data through neighboring utterance matching and pseudo-segmentation. Extensive experiments on two benchmark datasets (i.e., DialSeg711 and Doc2Dial) demonstrate that our method significantly outperforms the strong baseline methods. For reproducibility, we provide our code and data at:https://github.com/AlibabaResearch/DAMO-ConvAI/tree/main/dial-start.

View on arXivView PDF

Code Implementations(1)

AlibabaResearch/DAMO-ConvAIOfficial100%
MIT

DAMO-ConvAI: The official repository which contains the codebase for Alibaba DAMO Conversational AI.

1,602252CSS, MDXJul 15, 20229 months agoMIT
conversational-aideep-learningdialognatural-language-processing

Similar Papers

An Unsupervised Dialogue Topic Segmentation Model Based on Utterance Rewriting

89% match2024

Multi-Stage Coarse-to-Fine Contrastive Learning for Conversation Intent Induction

84% match2023

CTRLStruct: Dialogue Structure Learning for Open-Domain Response Generation

83% match1 repo92023

Multi-turn Dialogue Comprehension from a Topic-aware Perspective

83% match2023

SuperDialseg: A Large-scale Dataset for Supervised Dialogue Segmentation

82% match1 repo72023

Template-assisted Contrastive Learning of Task-oriented Dialogue Sentence Embeddings

82% match1 repo12023

Cite this paper

@article{gao2023unsupervised,
  title  = {Unsupervised Dialogue Topic Segmentation with Topic-aware Utterance Representation},
  author = {Haoyu Gao and Rui Wang and Ting-En Lin and Yuchuan Wu and Min Yang and Fei Huang and Yongbin Li},
  year   = {2023},
  eprint = {2305.02747},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2305.02747v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2305.02747.svg)](https://codeofpaper.com/paper/2305.02747)

Discussion