Sogang University researcher awarded fellowship for AI hardware innovation

Dr. Jin Huding (Kim Ho-jeong), a postdoctoral researcher at Sogang University's G-LAMP program, has secured a prestigious fellowship that will fuel groundbreaking research in artificial intelligence hardware. Selected for the 2026 Sejong Science Fellowship (Return & Attraction) program, funded by the Ministry of Science and ICT and managed by the National Research Foundation of Korea, Dr. Jin will receive 500 million won over five years to pursue her project titled 'Research on Next-Generation Artificial Intelligence Hardware Based on Self-Powered Multi-Modal Neuromorphic Sensing-Computing Integrated Architecture.'

The research addresses a critical challenge in current AI development: the inefficiency of existing hardware. Traditional AI systems, based on the Von Neumann architecture, separate sensing, computation, and actuation. This separation leads to significant energy consumption and delays when processing the vast amounts of sensory data required by autonomous systems, especially in environments like the Internet of Things (IoT) or wearable devices where constant sensor operation is necessary and external power sources are often impractical. Dr. Jin's proposed solution is a novel hardware architecture that integrates sensing and computation, enabling the system to generate signals and perform neuromorphic calculations using energy harvested directly from sensory stimuli, thus eliminating the need for external power.

This research pushes beyond existing studies on self-powered neuromorphic devices, which have largely focused on single sensory modalities. Dr. Jin's project emphasizes the fusion of multi-modal sensory inputs and cross-modal computation at the hardware level, a key aspect of human cognition. A notable example is the integration of taste (chemical) and vision (light) sensing, which has been challenging to implement. The goal is to create an integrated architecture capable of self-powered signal generation, neuromorphic synaptic operations, and cross-modal associative learning, even in environments where both chemical and light stimuli are present. This work holds significant academic and technological promise, potentially revolutionizing AI hardware by enabling ultra-low-power, autonomous operation for applications in edge computing, IoT, and autonomous systems.