Zhijie Yang

Zhijie Yang

Auto-generated profile, based on publicly available information

Zhijie Yang

National Innovation Institute of Defense Technology, Academy of Military Science, Beijing, China

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 February 2022 in IEEE Transactions on Circuits and Systems I: Regular Papers
Verified
Journal Article
Closed access
LWLei Wang
Lei Wang
ZYZhijie Yang
Zhijie Yang
SGShasha Guo
Shasha Guo
LQLianhua Qu
Lianhua Qu
XZXiangyu Zhang
Xiangyu Zhang
WXWeixia Xu
Weixia Xu
Neuromorphic processors have gained momentum recently due to their high energy efficiency in artificial intelligence applications compared to DNN accelerators. Most neuromorphic processors are executing SNNs (Spiking Neural Networks). Liquid State Machine (LSM), as the spiking version of reservoir computing, shows advantages and great potential in image classification, speech recognition, language translation, etc.. Comparing with other SNN models, LSM has the characteristics of easy to train and low resource utilization, which is suitable for low-power and resource-constrained edge computing scenarios. In this paper, we propose a novel design of a neuromorphic processor, LSMCore, aiming at LSM acceleration. LSMCore supports both training and inference of LSM. It consists of 256 input neurons, 1024 liquid neurons, and 1.31M synapses. Besides, multiple optimization techniques, including weight quantization for reducing storage, zero-skipping for decreasing dynamic sparsity, and mini-batch training are adopted in this processor. The experimental results show that the frequency of LSMCore achieves 400 MHz, the power is 4.9W and the area is 18.49 mm 2 with a 40nm library. Comparing with the baseline, LSMCore achieves up to $80.7\times $ ( $49.6\times $ ), $91.3\times $ ( $56.3\times $ ), and $83.1\times $ ( $56.8\times $ ) speedup on MNIST, N-MNIST, and Free Spoken Digital Dataset (FSDD) respectively for training (inference), while the accuracy of LSMCore on these three datasets are 96.8%, 97.6%, and 90% respectively.

19 February 2020 in IEEE Embedded Systems Letters
Verified
Journal Article
Closed access
YZYang Zhijie
Yang Zhijie
WLWang Lei
Wang Lei
LLLuo Li
Luo Li
LSLi Shiming
Li Shiming
GSGuo Shasha
Guo Shasha
WSWang Shuquan
Wang Shuquan
In recent years, convolutional neural networks (CNNs) have been widely used. However, their ever-increasing amount of parameters makes it challenging to train them with the GPUs, which is time and energy expensive. This has prompted researchers to turn their attention to training on more energy-efficient hardware. batch normalization (BN) layer has been widely used in various state-of-the-art CNNs for it is an indispensable layer in the acceleration of CNN training. As the amount of computation of the convolutional layer declines, its importance continues to increase. However, the traditional CNN training accelerators do not pay attention to the efficient hardware implementation of the BN layer. In this letter, we design an efficient CNN training architecture by using the systolic array. The processing element of the systolic array can support the BN functions both in the training process and the inference process. The BN function implemented is an improved, hardware-friendly BN algorithm, range batch normalization (RBN). The experimental results show that the implementation of RBN saves 10% hardware resources, reduces the power by 10.1%, and the delay by 4.6% on average. We implement the accelerator on the field programmable gate array VU440, and the power consumption of the its core computing engine is 8.9 W.

22 August 2023 in IEEE Transactions on Parallel and Distributed Systems
Verified
Journal Article
Closed access
ZYZhijie Yang
Zhijie Yang
LWLei Wang
Lei Wang
WSWei Shi
Wei Shi
YWYao Wang
Yao Wang
JTJunbo Tie
Junbo Tie
WXWeixia Xu
Weixia Xu
In recent years, neuromorphic processors are widely used in many scenarios, showing extreme energy efficiency over traditional architectures. However, almost all existing neuromorphic hardware are following the heterogeneous computing methodology without Instruction Set Architecture (ISA), leading to inflexibility in programming. In this paper, we first propose a RISC-V Neuromorphic Extension (RVNE) to enable fine-grained and flexible homogeneous programming for neuromorphic algorithms while utilizing SNN sparsity from different levels of granularity and computing flows. Based on RVNE, we next implement a neuromorphic micro-architecture that is tightly coupled to the CPU pipeline to accelerate neuromorphic computing. To demonstrate the proposed homogeneous neuromorphic architecture, we implement a prototype processor called NeuroRVcore based on RISC-V ISA and an open-source RISC-V core. The evaluation results show that RVNE achieves a 2.8 × −4.3 × reduction in code density compared with the general-purpose ISAs. Compared with the state-of-the-art neuromorphic processor, the proposed homogeneous computing reduces energy consumption by 3.4%−22.5% while enabling fine-grained and flexible homogeneous programming.

29 June 2025
Verified
Conference Paper
Open access
ZFZhigang Fang
Zhigang Fang
RCRenzhi Chen
Renzhi Chen
ZYZhijie Yang
Zhijie Yang
YGYang Guo
Yang Guo
HDHuadong Dai
Huadong Dai
LWLei Wang
Lei Wang

18 December 2023 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
Verified
Journal Article
Closed access
CXChao Xiao
Chao Xiao
XHXu He
Xu He
ZYZhijie Yang
Zhijie Yang
XXXun Xiao
Xun Xiao
YWYao Wang
Yao Wang
WXWeixia Xu
Weixia Xu
Neuromorphic processors have been designed as nonvon Neumann systems for energy-efficient spiking neural network (SNN) execution. Spiking convolutional neural networks (SCNNs), combining the advantage of convolutional neural network (CNN) and SNN, have been widely applied to vision tasks. However, as the scale of SCNNs increases, executing large-scale SCNNs on resource-constrained neuromorphic processor faces many challenges, including massive synapse pruning, caused by resource competition, execution performance degradation, etc. Addressing these problems, we propose an efficient approach to map large-scale SCNNs onto resource-constrined neuromorphic processor. The approach consists of three steps: 1) splitting; 2) partitioning; and 3) mapping. We explore three acyclic splitting strategies to divide large-scale SCNNs into subnetworks without cyclic dependency. Axon sharing is the guiding principle to partition subnetworks into multiple clusters. To obtain an optimal cluster-to-core mapping scheme, we use nondominated sorting genetic algorithm to collaboratively optimize two metrics. We evaluate our approach with eight realistic SCNN applications. The results show that compared with existing state-of-the-art methods, our approach significantly reduces the synapse pruning and accuracy loss, and increases the execution performance.

15 March 2021 in Electronics
Verified
Journal Article
Open access
PDF file
DDDong Ding
Dong Ding
LWLei Wang
Lei Wang
ZYZhijie Yang
Zhijie Yang
KHKai Hu
Kai Hu
HHHongjun He
Hongjun He
Analog Computing In Memory (ACIM) combines the advantages of both Compute In Memory (CIM) and analog computing, making it suitable for the design of energy-efficient hardware accelerators for computationally intensive DNN applications. However, their use will introduce hardware faults that decrease the accuracy of DNN. In this work, we take Sandwich-Ram as the real hardware example of ACIM and are the first to propose a fault injection and fault-aware training framework for it, named Analog Computing In Memory Simulator (ACIMS). Using this framework, we can simulate and repair the hardware faults of ACIM. The experimental results show that ACIMS can recover 91.0%, 93.7% and 89.8% of the DNN’s accuracy drop through retraining on the MNIST, SVHN and Cifar-10 datasets, respectively; moreover, their adjusted accuracy can reach 97.0%, 95.3% and 92.4%.