Semi-supervised new slot discovery with incremental clustering

Discovering new slots is critical to the success of dialogue systems. Most existing methods rely on automatic slot induction in an unsupervised fashion or perform domain adaptation across zero or few-shot scenarios. They have difficulties in providing high-quality supervised signals to learn cluster...

Full description

Saved in:
Bibliographic Details
Main Authors: WU, Yuxia, LIAO, Lizi, QIAN, Xuemin, CHUA, Tat-Seng
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2022
Subjects:
Online Access:https://ink.library.smu.edu.sg/sis_research/7736
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
id sg-smu-ink.sis_research-8739
record_format dspace
spelling sg-smu-ink.sis_research-87392023-01-10T02:00:04Z Semi-supervised new slot discovery with incremental clustering WU, Yuxia LIAO, Lizi QIAN, Xuemin CHUA, Tat-Seng Discovering new slots is critical to the success of dialogue systems. Most existing methods rely on automatic slot induction in an unsupervised fashion or perform domain adaptation across zero or few-shot scenarios. They have difficulties in providing high-quality supervised signals to learn clustering-friendly features, and are limited in effectively transferring the prior knowledge from known slots to new slots. In this work, we propose a Semi-supervised Incremental Clustering method (SIC), to discover new slots with the aid of existing linguistic annotation models and limited known slot data. Specifically, we harvest slot value candidates with NLP model cues and innovatively formulate the slot discovery task under an incremental clustering framework. The model gradually calibrates slot representations under the supervision of generated pseudo-labels, and automatically learns to terminate when no more salient slot remains. Our thorough evaluation on five public datasets demonstrates that the proposed method significantly outperforms state-of-theart models. 2022-12-11T08:00:00Z text https://ink.library.smu.edu.sg/sis_research/7736 Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University Incremental data Incremental cluster Artificial Intelligence and Robotics
institution Singapore Management University
building SMU Libraries
continent Asia
country Singapore
Singapore
content_provider SMU Libraries
collection InK@SMU
language English
topic Incremental data
Incremental cluster
Artificial Intelligence and Robotics
spellingShingle Incremental data
Incremental cluster
Artificial Intelligence and Robotics
WU, Yuxia
LIAO, Lizi
QIAN, Xuemin
CHUA, Tat-Seng
Semi-supervised new slot discovery with incremental clustering
description Discovering new slots is critical to the success of dialogue systems. Most existing methods rely on automatic slot induction in an unsupervised fashion or perform domain adaptation across zero or few-shot scenarios. They have difficulties in providing high-quality supervised signals to learn clustering-friendly features, and are limited in effectively transferring the prior knowledge from known slots to new slots. In this work, we propose a Semi-supervised Incremental Clustering method (SIC), to discover new slots with the aid of existing linguistic annotation models and limited known slot data. Specifically, we harvest slot value candidates with NLP model cues and innovatively formulate the slot discovery task under an incremental clustering framework. The model gradually calibrates slot representations under the supervision of generated pseudo-labels, and automatically learns to terminate when no more salient slot remains. Our thorough evaluation on five public datasets demonstrates that the proposed method significantly outperforms state-of-theart models.
format text
author WU, Yuxia
LIAO, Lizi
QIAN, Xuemin
CHUA, Tat-Seng
author_facet WU, Yuxia
LIAO, Lizi
QIAN, Xuemin
CHUA, Tat-Seng
author_sort WU, Yuxia
title Semi-supervised new slot discovery with incremental clustering
title_short Semi-supervised new slot discovery with incremental clustering
title_full Semi-supervised new slot discovery with incremental clustering
title_fullStr Semi-supervised new slot discovery with incremental clustering
title_full_unstemmed Semi-supervised new slot discovery with incremental clustering
title_sort semi-supervised new slot discovery with incremental clustering
publisher Institutional Knowledge at Singapore Management University
publishDate 2022
url https://ink.library.smu.edu.sg/sis_research/7736
_version_ 1770576423796342784