JSAI2024

Presentation information

Organized Session

Organized Session » OS-13

[2R6-OS-13a] OS-13

Wed. May 29, 2024 5:30 PM - 7:10 PM Room R (Room 51)

オーガナイザ:酒井 元気(日本大学)、岡田 将吾(北陸先端科学技術大学院大学)、湯浅 将英(湘南工科大学)、近藤 一晃(京都大学)、下西 慶(京都大学)

6:30 PM - 6:50 PM

[2R6-OS-13a-04] Modeling the adjusting of assumptions in the grounding

〇Takeru Amaya1, Tatsuya Yui1, Junya Morita1, Ryuichiro Higashinaka2, Yugo Takeuchi1 (1. Shizuoka University, 2. NTT)

Keywords:Grounding, Dialogue, Interactive Agent

In tasks requiring collaboration through dialogue, forming a shared understanding or "common ground" is key to solving problems. Conversations often use vague expressions that can apply to many instances or convey an overall impression, making them seem inefficient for building common ground. Yet, by examining a strategy that utilizes the relationships between subjects without detailed mention, we can see dialogues with ambiguous utterances as a way to align mutual expectations through an orderly strategy. This process is complex, needing constant updates to others' perceptions in real-time. By modeling this behavior, we can simulate how dialoguers form common ground through complementary speech and make internal estimations, aiding in developing dialogue agents that can produce more accurate expressions. This study analyzes how ambiguous expressions contribute to forming common ground and aligning mutual expectations in dialogue, focusing on the dynamic changes over time.

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