Sho Yatabe

About

Sho Yatabe is a researcher advancing efficient deep learning hardware, with a focus on approximate computing for convolutional neural network (CNN) accelerators. His key contributions lie in developing area-efficient and low-power architectures for edge AI, particularly targeting real-time robotic control and reduced network load. His most cited work, "Area-efficient Binary and Ternary CNN Accelerator using Random-forest-based Approximation" (2021), introduces a novel random-forest-based approximation layer unit (RFA-LU) that significantly reduces hardware footprint while maintaining inference accuracy. This work has garnered 4 citations, reflecting its niche but growing impact in the field of approximate computing for neural networks. Yatabe's research bridges the gap between algorithmic efficiency and practical hardware implementation, addressing critical challenges in deploying CNNs on resource-constrained devices. His innovations are particularly relevant for low-latency applications where traditional accelerators fall short, positioning him as a contributor to the next generation of compact, energy-efficient AI systems.

Research Focus

Key Achievements

Top Papers

1. 1
   Area-efficient Binary and Ternary CNN Accelerator using Random-forest-based Approximation

   4 citations · 2021

Key Collaborators

Contact & Links