JSAI2020

Presentation information

Organized Session

Organized Session » OS-16

[3P1-OS-16a] OS-16 (1)

Thu. Jun 11, 2020 9:00 AM - 10:40 AM Room P (jsai2020online-16)

佐々木 淳(AOI TYO Holdings株式会社)、阿部 明典(千葉大学)

10:20 AM - 10:40 AM

[3P1-OS-16a-04] A Bayesian approach to cognitive modeling of sarcasm comprehension based on the implicit display theory

〇Akira Utsumi1 (1. The University of Electro-Communications)

Keywords:irony, cognitive model

According to the implicit display theory, irony (and sarcasm) implicitly displays ironic environment (i.e., situational setting that motivates irony). The implicit display theory is compatible with, and can be modeled by, a Bayesian approach to cognition. The posterior probability of a given utterance to be sarcastic in a given context can be computed from a prior probability defined by the notion of ironic environment and the likelihood (of the utterance to be used with sarcastic intention) defined by the notion of implicit display. In this paper, the probability of texts to be interpreted sarcastically was automatically computed using sentiment analysis and compared with human sarcasm ratings on the same texts. As a result, the estimated probability of sarcasm was weakly but significantly correlated with human ratings (r=0.39). This suggests the psychological plausibility of the proposed cognitive model.

Authentication for paper PDF access

A password is required to view paper PDFs. If you are a registered participant, please log on the site from Participant Log In.
You could view the PDF with entering the PDF viewing password bellow.

Password