Yi Liu

Yi Liu


Research Focus & Metrics

Academic Career Length

9Years (2018-Present)

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

Representative Publications

View All Publications
16 September 2022 in Symmetry
Verified
Journal Article
Open access
PDF file
Changqing XuChangqing Xu
Changqing Xu
Yi LiuYi Liu
Yi Liu
Dongdong ChenDongdong Chen
Dongdong Chen
Yintang YangYintang Yang
Yintang Yang
Spiking neural networks (SNNs) can utilize spatio-temporal information and have the characteristic of energy efficiency, being a good alternative to deep neural networks (DNNs). The event-driven information processing means that SNNs can reduce the expensive computation of DNNs and save a great deal of energy consumption. However, high training and inference latency is a limitation of the development of deeper SNNs. SNNs usually need tens or even hundreds of time steps during the training and inference process, which causes not only an increase in latency but also excessive energy consumption. To overcome this problem, we propose a novel training method based on backpropagation (BP) for ultra-low-latency (1–2 time steps) SNNs with multi-threshold. In order to increase the information capacity of each spike, we introduce the multi-threshold Leaky Integrate and Fired (LIF) model. The experimental results show that our proposed method achieves average accuracy of 99.56%, 93.08%, and 87.90% on MNIST, FashionMNIST, and CIFAR10, respectively, with only two time steps. For the CIFAR10 dataset, our proposed method achieves 1.12% accuracy improvement over the previously reported directly trained SNNs with fewer time steps.

25 February 2023 in Symmetry
Verified
Journal Article
Open access
PDF file
XWxiaodong weng
xiaodong weng
Yi LiuYi Liu
Yi Liu
Changqing XuChangqing Xu
Changqing Xu
XLXiaoling Lin
Xiaoling Lin
LZLinjun Zhan
Linjun Zhan
Yintang YangYintang Yang
Yintang Yang
Network on chip (NoC) is a promising solution to the challenge of multi-core System-on-Chip (SoC) communication design. Application mapping is the first and most important step in the NoC synthesis flow, which determines most of the NoC design performance. NoC mapping has been confirmed as an NP-hard (Non-Polynomial hard) problem, which could not be solved in polynomial time. Various heuristic mapping algorithms have been applied to the mapping problem. However, the heuristic algorithm easily falls into a local optimal solution which causes performance loss. Additionally, regular topologies of NoC, such as the ring, torus, etc., may generate symmetric solutions in the NoC mapping process, which increase the performance loss. Machine learning involves data-driven methods to analyze trends, find relationships, and develop models to predict things based on datasets. In this paper, an NoC machine learning mapping algorithm is proposed to solve a mapping problem. A Low-complexity and no symmetry NoC mapping dataset is defined, and a data augmentation approach is proposed to build dataset. With the dataset defined, a multi-label machine learning is established. The simulation results have confirmed that the machine learning mapping algorithm is proposed have at least 99.6% model accuracy and an average of 96.3% mapping accuracy.

9 August 2022 in Symmetry
Verified
Journal Article
Open access
PDF file
Changqing XuChangqing Xu
Changqing Xu
Jiahao NingJiahao Ning
Jiahao Ning
Yi LiuYi Liu
Yi Liu
MLMintao Luo
Mintao Luo
Dongdong ChenDongdong Chen
Dongdong Chen
Yintang YangYintang Yang
Yintang Yang
Mapping optimization of network-on-chips (NoCs) for specific applications has become one of the most important keys of the SoC top-level design. However, the topology of NoC applied is usually regular topology, such as mesh, torus, etc., which may generate a large number of isomorphic solutions during the process of NoC mapping, which may reduce the convergence speed of mapping algorithms. In this paper, we proposed a generic-based hyper-heuristic algorithm named IRC-GHH for NoC mapping. To reduce the influence of isomorphic solutions, we analyzed the symmetry of NoC topology and proposed crossover operators based on the isomorphic solution to optimize the algorithm. We studied the situation of invalid crossovers and eliminated invalid iterations by adopting an isomorphic replacement crossover (IRC) strategy. To prevent the algorithm from falling into evolutionary stagnation in the late iteration, we introduce an adaptive mechanism to increase the usage frequency of the IRC operator automatically. Compared with GHH without IRC, the GHH with IRC can achieve, on average 15.25% communication energy reduction and 7.84% communication delay reduction.

11 May 2023 in Micromachines
Verified
Journal Article
Open access
PDF file
CWChen Wang
Chen Wang
Yi LiuYi Liu
Yi Liu
Changqing XuChangqing Xu
Changqing Xu
XLXinfang Liao
Xinfang Liao
Dongdong ChenDongdong Chen
Dongdong Chen
ZWZhenyu Wu
Zhenyu Wu
Power MOSFETs are found to be very vulnerable to single-event burnout (SEB) in space irradiation environments, and the military components generally require that devices could operate reliably as the temperature varies from 218 K to 423 K (−55 °C to 150 °C); thus, the temperature dependence of single-event burnout (SEB) in power MOSFETs should be investigated. Our simulation results showed that the Si power MOSFETs are more tolerant to SEB at a higher temperature at the lower LET (10 MeV∙cm2/mg) due to the decrease of the impact ionization rate, which is in good agreement with the previous research. However, the state of the parasitic BJT plays a primary role in the SEB failure mechanism when the LET value is greater than 40 MeV∙cm2/mg, which exhibits a completely different temperature dependence from that of 10 MeV∙cm2/mg. Results indicate that with the temperature increasing, the lower difficulty to turn on the parasitic BJT and the increasing current gain all make it easier to build up the regenerative feedback process responsible for SEB failure. As a result, the SEB susceptibility of power MOSFETs increases as ambient temperature increases when the LET value is greater than 40 MeV∙cm2/mg.

13 June 2025 in Electronics
Verified
Journal Article
Open access
PDF file
Jiajun HeJiajun He
Jiajun He
Yi LiuYi Liu
Yi Liu
Changqing XuChangqing Xu
Changqing Xu
XLXinfang Liao
Xinfang Liao
Yintang YangYintang Yang
Yintang Yang
Over the past decade, deep neural networks (DNNs) have revolutionized the fields of computer vision (CV) and natural language processing (NLP), achieving unprecedented performance across a variety of tasks. The Vision Transformer (ViT) has emerged as a powerful alternative to convolutional neural networks (CNNs), leveraging self-attention mechanisms to capture long-range dependencies and global context. Owing to their flexible architecture and scalability, ViTs have been widely adopted in safety-critical applications such as autonomous driving, where system reliability is paramount. However, ViTs’ reliability issues induced by soft errors in large-scale digital integrated circuits have generally been overlooked. In this paper, we present a fine-grained fault sensitivity analysis of ViT variants under bit-flip fault injections, focusing on different ViT models, transformer encoder layers, weight matrix types, and attention-head dimensions. Experimental results demonstrate that the first transformer encoder layer is susceptible to soft errors due to its essential role in local and global feature extraction. Moreover, in the middle and later layers, the Multi-Layer Perceptron (MLP) sub-blocks dominate the computational workload and significantly influence representation learning, making them critical points of vulnerability. These insights highlight key reliability bottlenecks in ViT architectures when deployed in error-prone environments.

17 February 2023 in Micromachines
Verified
Journal Article
Open access
PDF file
XWxiaodong weng
xiaodong weng
XLXiaoling Lin
Xiaoling Lin
Yi LiuYi Liu
Yi Liu
Changqing XuChangqing Xu
Changqing Xu
LZLinjun Zhan
Linjun Zhan
Yintang YangYintang Yang
Yintang Yang
Network on chip (NoC) is the main solution to the communication bandwidth of a multi-processor system on chip (MPSoC). NoC also brings more route requirements and is highly prone to errors caused by crosstalk. Crosstalk has become a major design problem in deep-submicron NoC communication design. Hence, a crosstalk error model and corresponding reliable system with error correction code (ECC) are required to make NoC communication reliable. In this paper, a reliability system evaluation model (RSE) of NoC communication with analysis from backend to frontend has been proposed. In the backend, a crosstalk error rate model (CER) is established with a three-wire RLC coupling model and timing constraints. The CER is used to establish functional relations between interconnect spacing, length and signal frequency, and test system reliability. In the frontend, a reliability system performance model (RSP) is established with a CER, reliability method cost and bandwidth. The RSE summarizes the frontend and backend model. In order to verify the RSE model, we propose a reliability system with a hybrid automatic repeat request technique (RSHARQ). Simulation demonstrates that the CER model is close to real circuit design. Through the CER and RSP model, the performance of RSHARQ could be simulated.