Shiori Kuramoto

Papers

Total Citations

H-Index

About

Shiori Kuramoto’s research lies at the intersection of robotics, neural network interpretability, and embodied cognition. Her most cited work, “Visualization of topographical internal representation of learning robots” (2020, 5 citations), pioneers methods to decode how neural network-controlled robots internalize their physical environments during learning. By visualizing the hidden layers of these networks, Kuramoto reveals how learned strategies map onto real-world topographies—a critical step toward making “black-box” robotic learning transparent. This work bridges the gap between abstract neural representations and tangible robotic behavior, offering insights for designing more adaptive, explainable autonomous systems. Though early in her career, Kuramoto’s focus on understanding the interplay between physical learning environments and internal network dynamics positions her as a rising voice in interpretable AI for robotics. Her contributions carry significant implications for fields ranging from developmental robotics to human-robot interaction, where trust and transparency are paramount. As she continues to explore how robots form internal models of their surroundings, Kuramoto is shaping a future where machines not only learn but also reveal the logic behind their actions.