Jin Huding, LAMP Postdoctoral Researcher at G-LAMP, Selected for the 2026 Sejong Science Fellowship (Return and Recruitment Track) of the National Research Foundation of Korea
Jin Huding, a LAMP postdoctoral researcher at this university’s G-LAMP, has been selected for the 2026 Sejong Science Fellowship (Return and Recruitment Track), a basic research initiative in science and technology organized by the Ministry of Science and ICT and the National Research Foundation of Korea. The research project, titled “Research on Next-Generation AI Hardware Based on a Self-Powered Multi-modal Neuromorphic Sensing-Computing Integrated Architecture,” will receive a total of 500 million KRW in funding over five years, from September 2026 to August 2031.
While demand for autonomous intelligent systems capable of independently perceiving and making decisions about their surroundings is growing, current AI hardware is based on the von Neumann architecture, in which sensors, computation, and actuation are separated. This structure inevitably leads to energy consumption and latency when processing large-scale sensory data. In particular, in environments where multiple sensor nodes must operate continuously, such as the Internet of Things (IoT), wearables, and robotics, the existing architecture faces fundamental limitations due to its reliance on external power and data transmission. To address this, Jin Huding proposed a “perception-computation integrated” hardware architecture that generates signals by harvesting energy directly from sensory stimuli without an external power source, while simultaneously performing neuromorphic computations.
In particular, this research goes beyond the limitations of previous studies on self-powered neuromorphic devices, which mostly focused on a single sense, and focuses on implementing multimodal sensory fusion and cross-modal computation, which are central to human cognition, at the hardware level. A representative example is the fusion of taste (chemical) and vision (optical) senses, an area where research has been relatively scarce due to the high difficulty of implementation. The goal is to implement an integrated architecture capable of self-powered signal generation, neuromorphic synaptic operation, and cross-modal associative learning even in environments where chemical and optical stimuli coexist.
This research holds significant academic and technical importance as it fundamentally transforms the existing AI hardware paradigm, in which sensory reception, computation, and energy are separated, and presents foundational technologies for next-generation AI hardware capable of autonomous operation even in ultra-low-power environments. The in-sensor structure, which fundamentally reduces energy consumption associated with data transfer, can be extended to power-constrained environments such as edge computing, artificial intelligence, the Internet of Things (IoT), and autonomous systems. It is also expected to be applied to intelligent sensor systems capable of situational awareness and decision-making, such as environmental monitoring, sensory modules for autonomous robots, and human-machine interfaces.
[SEO keyword]
Sejong Science Fellowship, Neuromorphic Hardware, AI Hardware
While demand for autonomous intelligent systems capable of independently perceiving and making decisions about their surroundings is growing, current AI hardware is based on the von Neumann architecture, in which sensors, computation, and actuation are separated. This structure inevitably leads to energy consumption and latency when processing large-scale sensory data. In particular, in environments where multiple sensor nodes must operate continuously, such as the Internet of Things (IoT), wearables, and robotics, the existing architecture faces fundamental limitations due to its reliance on external power and data transmission. To address this, Jin Huding proposed a “perception-computation integrated” hardware architecture that generates signals by harvesting energy directly from sensory stimuli without an external power source, while simultaneously performing neuromorphic computations.
In particular, this research goes beyond the limitations of previous studies on self-powered neuromorphic devices, which mostly focused on a single sense, and focuses on implementing multimodal sensory fusion and cross-modal computation, which are central to human cognition, at the hardware level. A representative example is the fusion of taste (chemical) and vision (optical) senses, an area where research has been relatively scarce due to the high difficulty of implementation. The goal is to implement an integrated architecture capable of self-powered signal generation, neuromorphic synaptic operation, and cross-modal associative learning even in environments where chemical and optical stimuli coexist.
This research holds significant academic and technical importance as it fundamentally transforms the existing AI hardware paradigm, in which sensory reception, computation, and energy are separated, and presents foundational technologies for next-generation AI hardware capable of autonomous operation even in ultra-low-power environments. The in-sensor structure, which fundamentally reduces energy consumption associated with data transfer, can be extended to power-constrained environments such as edge computing, artificial intelligence, the Internet of Things (IoT), and autonomous systems. It is also expected to be applied to intelligent sensor systems capable of situational awareness and decision-making, such as environmental monitoring, sensory modules for autonomous robots, and human-machine interfaces.
[SEO keyword]
Sejong Science Fellowship, Neuromorphic Hardware, AI Hardware