Learning influences fromword use in polylogue

Tomoharu Iwata, Shinji Watanabe

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We propose a probabilistic model for estimating influences among speakers from conversation data with multiple people. In conversations, people tend to mimic their companions' behavior depending on their level of trust. With the proposed model, we assume that the word use of a speaker depends on the word use of previous speakers as well as their own earlier word use and the general word distribution. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. Experiments on two meeting data sets in Japanese and in English demonstrate the effectiveness of the proposed method.

Original languageEnglish
Pages (from-to)3089-3092
Number of pages4
JournalUnknown Journal
Publication statusPublished - 2011
Externally publishedYes

Fingerprint

learning
conversation
Expectation-maximization Algorithm
Probabilistic Model
Experiments
Tend
Demonstrate
Experiment
estimating
Learning
Influence
Statistical Models
Model

Keywords

  • Conversation analysis
  • Influence
  • Latent variable model

ASJC Scopus subject areas

  • Language and Linguistics
  • Human-Computer Interaction
  • Signal Processing
  • Software
  • Modelling and Simulation

Cite this

Learning influences fromword use in polylogue. / Iwata, Tomoharu; Watanabe, Shinji.

In: Unknown Journal, 2011, p. 3089-3092.

Research output: Contribution to journalArticle

@article{65d1e318fd314830a113f756dccc47dc,
title = "Learning influences fromword use in polylogue",
abstract = "We propose a probabilistic model for estimating influences among speakers from conversation data with multiple people. In conversations, people tend to mimic their companions' behavior depending on their level of trust. With the proposed model, we assume that the word use of a speaker depends on the word use of previous speakers as well as their own earlier word use and the general word distribution. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. Experiments on two meeting data sets in Japanese and in English demonstrate the effectiveness of the proposed method.",
keywords = "Conversation analysis, Influence, Latent variable model",
author = "Tomoharu Iwata and Shinji Watanabe",
year = "2011",
language = "English",
pages = "3089--3092",
journal = "Nuclear Physics A",
issn = "0375-9474",
publisher = "Elsevier",

}

TY - JOUR

T1 - Learning influences fromword use in polylogue

AU - Iwata, Tomoharu

AU - Watanabe, Shinji

PY - 2011

Y1 - 2011

N2 - We propose a probabilistic model for estimating influences among speakers from conversation data with multiple people. In conversations, people tend to mimic their companions' behavior depending on their level of trust. With the proposed model, we assume that the word use of a speaker depends on the word use of previous speakers as well as their own earlier word use and the general word distribution. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. Experiments on two meeting data sets in Japanese and in English demonstrate the effectiveness of the proposed method.

AB - We propose a probabilistic model for estimating influences among speakers from conversation data with multiple people. In conversations, people tend to mimic their companions' behavior depending on their level of trust. With the proposed model, we assume that the word use of a speaker depends on the word use of previous speakers as well as their own earlier word use and the general word distribution. The influences can be efficiently estimated by using the expectation maximization (EM) algorithm. Experiments on two meeting data sets in Japanese and in English demonstrate the effectiveness of the proposed method.

KW - Conversation analysis

KW - Influence

KW - Latent variable model

UR - http://www.scopus.com/inward/record.url?scp=84865749839&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84865749839&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:84865749839

SP - 3089

EP - 3092

JO - Nuclear Physics A

JF - Nuclear Physics A

SN - 0375-9474

ER -