JSAI2020

Presentation information

Interactive Session

[4Rin1] Interactive 2

Fri. Jun 12, 2020 9:00 AM - 10:40 AM Room R01 (jsai2020online-2-33)

[4Rin1-25] Cognitive map and subjective time compression generated by Hierarchical RNN

〇Kohei Homma1, Norihiro Maruyama1, Atsushi Masumori1, Takashi Ikegami1 (1.The University of Tokyo, Komaba)

Keywords:Cognitive map, place cell, Artificial life

Animals develop cognitive maps, which are internal model of the external environment, to understand the own position and spatial characteristics of the environment, with multiple sense inputs.
Hippocampal place cell and grid cells are thought to play an important role in development this cognitive map. However, the mechanism remains not understood.Previous studies show that cognitive map self-organized in Higher-level layer of hierarchical RNN by predictive learning involving integrated vision and motion.
In addition, study using VAE-GAN with only visual information show that temporal acceleration/deceleration is observed during action on the generated cognitive map, and the relationship with subjective time compression has been discussed. Therefore, I modified previous model and involve vision and motion, and discuss the generation of cognitive maps and subjective 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