Professors Kim Jang-saeng, Kim Sang-wan, and Kim Si-hyun's Research Teams Selected for 2026 PIM AI Semiconductor Core Technology Development (Device) Project
A joint research team at the University, consisting of Professors Kim Jang-saeng, Kim Sang-wan, and Kim Si-hyun of the Departments of Electronic Engineering, System Semiconductor Engineering, and Semiconductor Engineering, has been selected for the 2026 PIM AI Semiconductor Core Technology Development (Device) Project, organized by the Ministry of Science and ICT and the National Research Foundation of Korea. The research project is titled “Bit-Weighted 3D IGZO FeNAND-Based Ultra-High-Dimensional In-Sensor Reservoir Computing Platform” and will receive a total of 600 million won in research funding over a period of two years and nine months, from April 2026 to December 2028.
With the recent proliferation of edge AI and on-device AI, in-sensor computing, which processes data directly at the sensor level, and PIM (Processing-In-Memory) technology, which performs computations within memory, are emerging as the next-generation semiconductor paradigms. In particular, in vision sensor-based AI applications, latency and energy consumption caused by data movement act as major bottlenecks, and the need for ultra-low-power, high-efficiency hardware architectures to address these issues is growing significantly. To overcome these limitations, this research aims to implement an ultra-high-dimensional in-sensor reservoir computing platform that integrates sensors, memory, and computation by utilizing a vertically stacked 3D FeNAND structure based on ferroelectric (HZO) and oxide semiconductor (IGZO) materials. Furthermore, the study proposes a novel computing architecture utilizing multiple masks to simultaneously address the integration limitations and state overlap issues inherent in existing reservoir computing. Through this R&D project, the research team will conduct full-cycle research spanning materials, devices, and systems.
This research is significant in that it fundamentally alleviates the data movement bottlenecks inherent in existing von Neumann architectures and presents a next-generation AI semiconductor architecture capable of processing high-dimensional time-series data at the sensor level with ultra-low power consumption. Furthermore, by introducing a new paradigm that enables efficient scaling of computational dimensions through the concept of multi-mask-based reservoir computing, it is expected to contribute to securing domestic technological competitiveness and industrial expansion in the next-generation semiconductor field.
[Keyword]
PIM AI Semiconductor, In-Sensor Reservoir Computing, Next-Generation Semiconductor Technology
[Summary]
A joint research team has been selected for the 2026 PIM AI Semiconductor Core Technology Development project to develop an ultra-high-dimensional computing platform. By utilizing 3D FeNAND structures, this study aims to advance In-Sensor Reservoir Computing and establish a new paradigm in Next-Generation Semiconductor Technology for efficient data processing.
With the recent proliferation of edge AI and on-device AI, in-sensor computing, which processes data directly at the sensor level, and PIM (Processing-In-Memory) technology, which performs computations within memory, are emerging as the next-generation semiconductor paradigms. In particular, in vision sensor-based AI applications, latency and energy consumption caused by data movement act as major bottlenecks, and the need for ultra-low-power, high-efficiency hardware architectures to address these issues is growing significantly. To overcome these limitations, this research aims to implement an ultra-high-dimensional in-sensor reservoir computing platform that integrates sensors, memory, and computation by utilizing a vertically stacked 3D FeNAND structure based on ferroelectric (HZO) and oxide semiconductor (IGZO) materials. Furthermore, the study proposes a novel computing architecture utilizing multiple masks to simultaneously address the integration limitations and state overlap issues inherent in existing reservoir computing. Through this R&D project, the research team will conduct full-cycle research spanning materials, devices, and systems.
This research is significant in that it fundamentally alleviates the data movement bottlenecks inherent in existing von Neumann architectures and presents a next-generation AI semiconductor architecture capable of processing high-dimensional time-series data at the sensor level with ultra-low power consumption. Furthermore, by introducing a new paradigm that enables efficient scaling of computational dimensions through the concept of multi-mask-based reservoir computing, it is expected to contribute to securing domestic technological competitiveness and industrial expansion in the next-generation semiconductor field.
[Keyword]
PIM AI Semiconductor, In-Sensor Reservoir Computing, Next-Generation Semiconductor Technology
[Summary]
A joint research team has been selected for the 2026 PIM AI Semiconductor Core Technology Development project to develop an ultra-high-dimensional computing platform. By utilizing 3D FeNAND structures, this study aims to advance In-Sensor Reservoir Computing and establish a new paradigm in Next-Generation Semiconductor Technology for efficient data processing.