activity
20192022
most citedNAND-SPIN-Based Processing-in-MRAM Architecture for Convolutional Neural Network Acceleration

20 citations · 49 across the 5 of their papers we have counts for

collaborators

6 papers

cs.AR202220 cited

NAND-SPIN-Based Processing-in-MRAM Architecture for Convolutional Neural Network Acceleration

Yinglin Zhao, Jianlei Yang, Bing Li +7

The performance and efficiency of running large-scale datasets on traditional computing systems exhibit critical bottlenecks due to the existing "power wall" and "memory wall" prob…

cs.LG202110 cited

FedSkel: Efficient Federated Learning on Heterogeneous Systems with Skeleton Gradients Update

Junyu Luo, Jianlei Yang, Xucheng Ye +2

Federated learning aims to protect users' privacy while performing data analysis from different participants. However, it is challenging to guarantee the training efficiency on het…

cs.AR202119 cited

S2Engine: A Novel Systolic Architecture for Sparse Convolutional Neural Networks

Jianlei Yang, Wenzhi Fu, Xingzhou Cheng +3

Convolutional neural networks (CNNs) have achieved great success in performing cognitive tasks. However, execution of CNNs requires a large amount of computing resources and genera…

cs.CL2021

RoSearch: Search for Robust Student Architectures When Distilling Pre-trained Language Models

Xin Guo, Jianlei Yang, Haoyi Zhou +2

Pre-trained language models achieve outstanding performance in NLP tasks. Various knowledge distillation methods have been proposed to reduce the heavy computation and storage requ…

cs.CV2020

SparseTrain: Exploiting Dataflow Sparsity for Efficient Convolutional Neural Networks Training

Pengcheng Dai, Jianlei Yang, Xucheng Ye +5

Training Convolutional Neural Networks (CNNs) usually requires a large number of computational resources. In this paper, \textit{SparseTrain} is proposed to accelerate CNN training…

cs.LG2019

Accelerating CNN Training by Pruning Activation Gradients

Xucheng Ye, Pengcheng Dai, Junyu Luo +4

Sparsification is an efficient approach to accelerate CNN inference, but it is challenging to take advantage of sparsity in training procedure because the involved gradients are dy…