Influence relation estimation based on lexical entrainment in conversation

Tomoharu Iwata, Shinji Watanabe

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

In conversations, people tend to mimic their companions' behavior depending on their level of trust. This phenomenon is known as entrainment. We propose a probabilistic model for estimating influences among speakers from conversation data involving multiple people by modeling lexical entrainment. The proposed model estimates word use as a function of the weighted sum of the earlier word use of other speakers. The weights represent influences between speakers. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. We also develop its online inference procedures for sequentially modeling the dynamics of influence relations. Experiments performed on two meeting data sets one in Japanese and one in English demonstrate the effectiveness of the proposed method.

Original languageEnglish
Pages (from-to)329-339
Number of pages11
JournalSpeech Communication
Volume55
Issue number2
DOIs
Publication statusPublished - 2013 Feb
Externally publishedYes

Fingerprint

Entrainment
conversation
Experiments
Expectation-maximization Algorithm
Weighted Sums
Modeling
Probabilistic Model
Tend
Influence
Statistical Models
experiment
Estimate
Demonstrate
Experiment

Keywords

  • Conversation analysis
  • Entrainment
  • Influence
  • Latent variable model

ASJC Scopus subject areas

  • Language and Linguistics
  • Linguistics and Language
  • Communication
  • Software
  • Computer Vision and Pattern Recognition
  • Computer Science Applications
  • Modelling and Simulation

Cite this

Influence relation estimation based on lexical entrainment in conversation. / Iwata, Tomoharu; Watanabe, Shinji.

In: Speech Communication, Vol. 55, No. 2, 02.2013, p. 329-339.

Research output: Contribution to journalArticle

@article{3ea0b1e1bb0e40328e1dcddd88c50d51,
title = "Influence relation estimation based on lexical entrainment in conversation",
abstract = "In conversations, people tend to mimic their companions' behavior depending on their level of trust. This phenomenon is known as entrainment. We propose a probabilistic model for estimating influences among speakers from conversation data involving multiple people by modeling lexical entrainment. The proposed model estimates word use as a function of the weighted sum of the earlier word use of other speakers. The weights represent influences between speakers. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. We also develop its online inference procedures for sequentially modeling the dynamics of influence relations. Experiments performed on two meeting data sets one in Japanese and one in English demonstrate the effectiveness of the proposed method.",
keywords = "Conversation analysis, Entrainment, Influence, Latent variable model",
author = "Tomoharu Iwata and Shinji Watanabe",
year = "2013",
month = "2",
doi = "10.1016/j.specom.2012.08.012",
language = "English",
volume = "55",
pages = "329--339",
journal = "Speech Communication",
issn = "0167-6393",
publisher = "Elsevier",
number = "2",

}

TY - JOUR

T1 - Influence relation estimation based on lexical entrainment in conversation

AU - Iwata, Tomoharu

AU - Watanabe, Shinji

PY - 2013/2

Y1 - 2013/2

N2 - In conversations, people tend to mimic their companions' behavior depending on their level of trust. This phenomenon is known as entrainment. We propose a probabilistic model for estimating influences among speakers from conversation data involving multiple people by modeling lexical entrainment. The proposed model estimates word use as a function of the weighted sum of the earlier word use of other speakers. The weights represent influences between speakers. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. We also develop its online inference procedures for sequentially modeling the dynamics of influence relations. Experiments performed on two meeting data sets one in Japanese and one in English demonstrate the effectiveness of the proposed method.

AB - In conversations, people tend to mimic their companions' behavior depending on their level of trust. This phenomenon is known as entrainment. We propose a probabilistic model for estimating influences among speakers from conversation data involving multiple people by modeling lexical entrainment. The proposed model estimates word use as a function of the weighted sum of the earlier word use of other speakers. The weights represent influences between speakers. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. We also develop its online inference procedures for sequentially modeling the dynamics of influence relations. Experiments performed on two meeting data sets one in Japanese and one in English demonstrate the effectiveness of the proposed method.

KW - Conversation analysis

KW - Entrainment

KW - Influence

KW - Latent variable model

UR - http://www.scopus.com/inward/record.url?scp=84870244007&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84870244007&partnerID=8YFLogxK

U2 - 10.1016/j.specom.2012.08.012

DO - 10.1016/j.specom.2012.08.012

M3 - Article

AN - SCOPUS:84870244007

VL - 55

SP - 329

EP - 339

JO - Speech Communication

JF - Speech Communication

SN - 0167-6393

IS - 2

ER -