Profile photo of Yuchang Huang

Yuchang Huang

Master's Student in Information Science and Engineering

Computing Architecture Laboratory
Nara Institute of Science and Technology, Japan

My research focuses on spiking neural networks, efficient Transformer architectures, neuromorphic computing, and surface electromyography-based human gesture recognition. I am particularly interested in developing energy-efficient artificial intelligence systems for biosignal understanding.

About Me

I am a master's student at the Nara Institute of Science and Technology (NAIST) and a member of the Computing Architecture Laboratory. My current research investigates efficient deep learning architectures for biosignal processing, with an emphasis on spiking neural networks and sEMG-based gesture recognition.

Before joining NAIST, I received my bachelor's degree from Yunnan University. My broader research goal is to bridge high-performance representation learning and energy-efficient neuromorphic computation.

Research Interests

Spiking Neural Networks

Event-driven neural computation, spike encoding, firing-rate control, and efficient temporal modeling.

Efficient Transformers

Lightweight attention mechanisms and efficient architectures for edge and neuromorphic systems.

Biosignal Understanding

Surface electromyography, human gesture recognition, and physiological time-series analysis.

Neuromorphic Computing

Low-power artificial intelligence and hardware-aware neural network design.

Publications

STGesture: A Topology-Aware Spiking Transformer for Sparse sEMG-Based Gesture Recognition

Yuchang Huang, Yirong Kan, and Yasuhiro Nakashima

Manuscript in preparation.

More publications will be added here.

This section will be updated as new research results become available.

Research Projects

STGesture

A topology-aware Spiking Transformer framework for efficient sparse sEMG gesture recognition.

Spiking Transformer sEMG Gesture Recognition

Topology-Aware Spike Encoding

Spike encoding methods designed to preserve local electrode relationships and stabilize neural activity.

Spike Encoding Electrode Topology

Local-Global Spike Attention

Lightweight local and global spike interaction mechanisms for reducing the computational cost of attention.

Attention Energy Efficiency

Neuromorphic Biosignal Processing

Energy-efficient neural architectures for processing temporal physiological signals on edge devices.

Neuromorphic AI Edge Computing

Education

2025 – Present

Nara Institute of Science and Technology

Master's Student, Information Science and Engineering

Computing Architecture Laboratory, Nara, Japan

2021 – 2025

Yunnan University

Bachelor's Degree

Kunming, China

Contact

Email: yuchang.huang.yf6@naist.ac.jp

GitHub: github.com/StubbornDuck

Affiliation: Computing Architecture Laboratory, Nara Institute of Science and Technology

Location: Nara, Japan