A humanoid locomotion approach based on a GPT-style multi-humanoid-task policy. Leveraging the general-purpose reasoning capabilities of large language models, the method realizes …
We propose a task-aware, robot-affordance-centric diffusion policy. By integrating the notion of affordances into a diffusion model, the proposed method enables more adaptive …
This work simulates the systems model of creativity using generative agents. By computationally reproducing human creative processes, we explore how creativity emerges.
We propose a multi-robot task planning method that integrates Large Language Models (LLMs) with linear programming and dependency graphs. By combining the natural-language …
A comprehensive survey of data-driven motion planning using deep neural networks, reinforcement learning, and large language models. We systematically organize state-of-the-art …
We propose a framework for explanation generation that aims to improve the reliability of autonomous robots. By presenting the decision-making process in a human-understandable …
Based on the framework of active inference, we propose an energy-minimization approach to multimodal affective human-robot interaction. The proposed mechanism enables a robot to …
This work computationally models how multimodal emotion perception (vision, audition, and touch) develops by leveraging tactile dominance and perceptual improvement. We reproduce …