Multimodal Negotiation Corpus with Various Subjective Assessments for Social-Psychological Outcome Prediction from Non-Verbal Cues

Proceedings of the 13th Conference on Language Resources and Evaluation (LREC 2022), pages 6794–6801 Marseille, 20-25 June 2022 © European Language Resources Association (ELRA), licensed under CC-BY-NC-4.0 6794 Multimodal Negotiation Corpus with Various Subjective Assessments for Social-Psychological Outcome Prediction from Non-Verbal Cues Nobuaktsu Hojo, Satoshi Kobashikawa, Saki Mizuno, Ryo Masumura NTT Computer & Science Laboratories 1-1 Hikarinooka, Yokosuka-shi, Kanagawa-ken 239-0847 nobukatsu.hojo.cd@hco.ntt.co.jp Abstract This study investigates social-psychological negotiation-outcome prediction (SPNOP), a novel task for estimating various sub- jective evaluation scores of negotiation, such as satisfaction and trust, from negotiation dialogue data. To investigate SPNOP, a corpus with various psychological measurements is beneficial because the interaction process of negotiation relates to many aspects of psychology. However, current negotiation corpora only include information related to objective outcomes or a single aspect of psychology. In addition, most use the “laboratory setting” that uses non-skilled negotiators and over simplified nego- tiation scenarios. There is a concern that such a gap with actual negotiation will intrinsically affect the behavior and psychology of negotiators in the corpus, which can degrade the performance of models trained from the corpus in real situations. Therefore, we created a negotiation corpus with three features; 1) was assessed with various psychological measurements, 2) used skilled negotiators, and 3) used scenarios of context-rich negotiation. We recorded video and audio of negotiations in Japanese to investigate SPNOP in the context of social signal processing. Experimental results indicate that social-psychological outcomes can be effectively estimated from multimodal information. Keywords: business negotiation, social-psychological outcomes, multimodal interaction analysis 1. Introduction Due to the impact of COVID-19, business negotia- tions are shifting from face-to-face to online. Since it is relatively easy to store data of online conversa- tions and for machines to intervene in such interac- tions, negotiation-related systems, such as negotiation- training systems (Greco and Murgia, 2007; Ding et al., 2017; Melzer, 2019) and negotiation-support sys- tems (Kersten, 1989; Carbonneau et al., 2016; Jonker et al., 2016; Zhou et al., 2019), will be more accessi- ble. As a fundamental technology for these systems, automatic negotiation-outcome prediction (Curhan and Pentland, 2007; Li et al., 2015) is important. This study investigated negotiation-outcome prediction from on- line negotiation data. Previous studies have shown that negotiation out- comes can be predicted using verbal (Zhao et al., 2018; Zhou et al., 2019) and non-verbal (Curhan and Pentland, 2007; Li et al., 2015) features ex- tracted from negotiation dialogues. In most cases, the term “negotiation outcome” refers to an economic outcome, i.e., the number of points earned by the participants’ joint decision making. Previous nego- tiation research argues that social-psychological out- comes (Thompson, 1990) are also important, espe- cially for long-term benefits (Barley, 1991; Oliver et al., 1994). Experimental evidence has also been pro- vided that trust, justice, and satisfaction lead to eco- nomic outcomes in the next round (Curhan et al., 2010). Therefore, we consider it important to predict social- psychological outcome from negotiation data. We call this task social-psychological negotiation-outcome prediction (SPNOP). The interaction process of negotiation is related to various …

Content truncated to 500 words