Research
3D Silicon Brain Architecture
We develop hardware for neuromorphic computing that mimics the behavior of biological brains. Here, the main target is to port the silicon brains into ASIC/FPGA and train them with bio-plausible learning approaches.
We have developed a LIF neuron model with STDP learning. Neurons are clustered and placed in a node of a 3D-NoC system that supports both unicast and multicast communications.
Furthermore, offline training has been investigating to deploy complex applications. We also plan to support the fault-tolerance feature for our 3D silicon brain.
Selected publications:
Mark Ogbodo, Khanh N. Dang, Abderazek Ben Abdallah, ‘On the Design of a Fault-tolerant Scalable Three Dimensional NoC-based Digital Neuromorphic System with On-chip Learning’’, IEEE Access, (accepted)
‘‘MigSpike: A Migration Based Algorithm and Architecture for Scalable Robust Neuromorphic Systems’’ (under major revision).
Duy-Anh Nguyen, Xuan-Tu Tran, Khanh N. Dang, and Francesca Iacopi, ‘‘A lightweight Max-Pooling method and architecture for Deep Spiking Convolutional Neural Networks’’, 2020 IEEE Asia Pacific Conference on Circuits and Systems (APCCAS), Dec. 8-10, 2020
Mark Ogbodo, The Vu, Khanh N. Dang and Abderazek Abdallah, ‘‘Light-weight Spiking Neuron Processing Core for Large-scale 3D-NoC based Spiking Neural Network Processing Systems’’, 2020 IEEE International Conference on Big Data and Smart Computing (BigComp), Feb. 19-22, 2020
Khanh N. Dang and Abderazek Ben Abdallah “An Efficient Software-Hardware Design Framework for Spiking Neural Network Systems”, 2019 International Conference on Internet of Things, Embedded Systems and Communications (IINTEC) (accepted).
Patent:
A. Ben Abdallah, Huakun Huang, Khanh N. Dang, Jiangning Song, ‘‘ＡＩプロセッサ (AI Processor)’’, 特願2020-194733, Japan patent, (patent filed)