<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Research Topics | Takato Horii | Osaka University</title><link>https://www.takatohorii.jp/en/research/</link><atom:link href="https://www.takatohorii.jp/en/research/index.xml" rel="self" type="application/rss+xml"/><description>Research Topics</description><generator>HugoBlox Kit (https://hugoblox.com)</generator><language>en-us</language><lastBuildDate>Sun, 19 May 2024 00:00:00 +0000</lastBuildDate><image><url>https://www.takatohorii.jp/media/icon_hu_da05098ef60dc2e7.png</url><title>Research Topics</title><link>https://www.takatohorii.jp/en/research/</link></image><item><title>Computational Modeling of Emotional Development</title><link>https://www.takatohorii.jp/en/research/emotion-development/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><guid>https://www.takatohorii.jp/en/research/emotion-development/</guid><description>&lt;p&gt;This line of research computationally models how human emotions develop through multimodal perception (vision, audition, touch). We focus on tactile dominance in infant-caregiver interaction and reproduce the developmental process of emotion perception using probabilistic generative models. We also investigate affective human-robot interaction based on an energy minimization framework grounded in Active Inference.&lt;/p&gt;</description></item><item><title>Constructive Study of Qualia and Consciousness</title><link>https://www.takatohorii.jp/en/research/qualia-consciousness/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><guid>https://www.takatohorii.jp/en/research/qualia-consciousness/</guid><description>&lt;p&gt;Can robots have emotional qualia? Addressing this question, we work on elucidating the correspondence between qualia structures and information structures. As part of the Transformative Research Areas (A) project on &amp;ldquo;Qualia Structurology,&amp;rdquo; we are building neural circuit models of subjective experience and frameworks based on probabilistic generative models. Joint research with IRCN at The University of Tokyo is also underway.&lt;/p&gt;</description></item><item><title>Human-Robot Interaction</title><link>https://www.takatohorii.jp/en/research/hri/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><guid>https://www.takatohorii.jp/en/research/hri/</guid><description>&lt;p&gt;Research on human-robot interaction aimed at the coexistence of humans and robots, including dialogue robots in social contexts, mental health support, and communication support. We address affective communication through touch and explainable decision-making frameworks based on graph-structured world models that ensure transparency of robotic decisions. Conducted as part of the Social Robotics Group (Yoshikawa Lab).&lt;/p&gt;</description></item><item><title>Robot Learning and LLM-Based Action Planning</title><link>https://www.takatohorii.jp/en/research/robot-learning/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><guid>https://www.takatohorii.jp/en/research/robot-learning/</guid><description>&lt;p&gt;Research on autonomous action planning and motion generation for robots using imitation learning, reinforcement learning, and Large Language Models (LLMs). We develop state-of-the-art methods including multi-robot task planning combining LLMs and linear programming (LiP-LLM), affordance-centric diffusion policies (TARAD), and humanoid locomotion (LocoGPT).&lt;/p&gt;</description></item><item><title>Symbol Emergence Robotics</title><link>https://www.takatohorii.jp/en/research/symbol-emergence/</link><pubDate>Mon, 01 Jan 2024 00:00:00 +0000</pubDate><guid>https://www.takatohorii.jp/en/research/symbol-emergence/</guid><description>&lt;p&gt;Research exploring how symbols and language can emerge through embodied robotic interaction. We work on integrated cognitive architectures of action and language, emergent communication through decentralized collective world models, and the theoretical construction of multi-timescale cognitive systems (System 0/1/2/3).&lt;/p&gt;</description></item></channel></rss>