Entrainment in Multi-Party Spoken Dialogues at Multiple Linguistic Levels

Rahimi, Zahra; Kumar, Anish; Litman, Diane; Paletz, Susannah B. F.; Yu, Mingzhi

doi:10.21437/interspeech.2017-1568

Cited by 27 publications

(51 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Predicting Social Outcomes: Our first task examines how the NW, PW, and GDW measures of acoustic-prosodic convergence (independent variables) relate to the social outcome measures (dependent variables) from Section 3. This is similar to prior studies which have evaluated convergence in terms of predicting outcomes Lee et al, 2011;Rahimi et al, 2017a). We hypothesize that the group-dynamic weighted convergence measure will outperform the nonweighted and participation-based measures.…”

Section: Experiments and Discussionsupporting

confidence: 86%

Weighting Model Based on Group Dynamics to Measure Convergence in Multi-party Dialogue

Rahimi¹,

Litman²

2018

Proceedings of the 19th Annual SIGdial Meeting on Discourse and Dialogue

Self Cite

View full text Add to dashboard Cite

show abstract

Section: Experiments and Discussionsupporting

confidence: 86%

Weighting Model Based on Group Dynamics to Measure Convergence in Multi-party Dialogue

Rahimi¹,

Litman²

2018

Proceedings of the 19th Annual SIGdial Meeting on Discourse and Dialogue

Self Cite

View full text Add to dashboard Cite

show abstract

“…So longer meetings tend to have lower values for mean_density_change and higher ratings for information use. Increases in network density may also relate to linguistic entrainment or alignment [24,25], wherein group members begin to use similar language as each other through the conversation.…”

Section: Resultsmentioning

confidence: 99%

Information Processing and Overload in Group Conversation: A Graph-Based Prediction Model

Murray

2019

MTI

View full text Add to dashboard Cite

Based on analyzing verbal and nonverbal features of small group conversations in a task-based scenario, this work focuses on automatic detection of group member perceptions about how well they are making use of available information, and whether they are experiencing information overload. Both the verbal and nonverbal features are derived from graph-based social network representations of the group interaction. For the task of predicting the information use ratings, a predictive model using random forests with verbal and nonverbal features significantly outperforms baselines in which the mean or median values of the training data are predicted, as well as significantly outperforming a linear regression baseline. For the task of predicting information overload ratings, the multimodal random forests model again outperforms all other models, including significant improvement over linear regression and gradient boosting models. However, on that task the best model is not significantly better than the mean and median baselines. For both tasks, we analyze performance using the full multimodal feature set versus using only linguistic features or only turn-taking features. While utilizing the full feature set yields the best performance in terms of mean squared error (MSE), there are no statistically significant differences, and using only linguistic features gives comparable performance. We provide a detailed analysis of the individual features that are most useful for each task. Beyond the immediate prediction tasks, our more general goal is to represent conversational interaction in such a way that yields a small number of features capturing the group interaction in an easily interpretable manner. The proposed approach is relevant to many other group prediction tasks as well, and is distinct from both classical natural language processing (NLP) as well as more current deep learning/artificial neural network approaches.

show abstract

“…Yu, 2013;Yu et al, 2013), but rather must be specific to particular characteristics. Little previous work has tested correlations across characteristics; the work that has looked for them has generally found no correlations (Bilous & Krauss, 1988;Pardo et al, 2012;Sanker, 2015;Weise & Levitan, 2018), with the occasional correlation that can probably be attributed to physical relationships between characteristics (Cohen Priva & Sanker, 2018) or repeated measures producing a false positive (Sanker, 2015;Rahimi et al, 2017). In studies with a relatively small number of participants, a null result for individual tendencies can be difficult to interpret; there has been no previous large-scale comparison across characteristics, controlling for effects of particular conversations.…”

Section: Speaker Effects In Convergencementioning

confidence: 99%

Natural Leaders: Some interlocutors elicit greater convergence across conversations and across characteristics

Priva¹,

Sanker²

2020

Preprint

View full text Add to dashboard Cite

Are there natural followers in human communication, speakers who consistently converge more than others? Similarly, are there natural leaders, speakers with whom others converge more? Are such tendencies consistent across different linguistic characteristics? We use the Switchboard Corpus, a large collection of telephone conversations, to perform a large-scale study of convergence of speakers in multiple conversations with different interlocutors, across six linguistic characteristics. Having data from each speaker in several conversations makes it possible to investigate whether there are individual differences in likelihood to converge, among speakers (more or less likely to converge), and among interlocutors (more or less likely to elicit convergence). We only find evidence for individual differences by interlocutor, not by speaker; this shows that there are natural leaders, who elicit greater degrees of convergence than others, across different characteristics and different conversations. The absence of a similar finding for natural followers, speakers who converge more than others, suggests that the role of social aspects in mediating convergence is stronger than that of putative individual differences in propensity to converge, or that such propensities are characteristic-specific.

show abstract

Entrainment in Multi-Party Spoken Dialogues at Multiple Linguistic Levels

Cited by 27 publications

References 26 publications

Weighting Model Based on Group Dynamics to Measure Convergence in Multi-party Dialogue

Weighting Model Based on Group Dynamics to Measure Convergence in Multi-party Dialogue

Information Processing and Overload in Group Conversation: A Graph-Based Prediction Model

Natural Leaders: Some interlocutors elicit greater convergence across conversations and across characteristics

Contact Info

Product

Resources

About