JSAI2025

Presentation information

Organized Session

Organized Session » OS-41

[1B5-OS-41c] OS-41

Tue. May 27, 2025 5:40 PM - 7:20 PM Room B (Small hall)

オーガナイザ:鈴木 雅大(東京大学),岩澤 有祐(東京大学),河野 慎(東京大学),熊谷 亘(オムロンサイニックエックス),松嶋 達也(東京大学),Paavo Parmas(東京大学),谷口 尚平(東京大学)

7:00 PM - 7:20 PM

[1B5-OS-41c-05] Perspective for Generative Emergent Communication

Is a large-scale language model a collective world model?

〇TADAHIRO TANIGUCHI1,4, Ryo Ueda2, Tomoaki Nakamura3, Masahiro Suzuki2, Akira Taniguchi4 (1. Kyoto University, 2. University of Tokyo, 3. The University of Electro-Communications, 4. Ritsumeikan University)

Keywords:emergent communication, world model, large language model, generative model, symbol emergence

This study proposes a unified theoretical framework called "generative emergent communication" (generative EmCom) that connects emergent communication, world models, and large language models from the perspective of collective predictive coding. This framework not only allows us to understand symbol emergence as (decentralized) representation learning of external symbols within a machine learning framework but also enables us to interpret large language models as collective world models that integrate experiences from multiple agents. This research provides a novel perspective linking world models and symbol emergence, while discussing future research directions.

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