Low-latency federated learning with DNN partition in distributed industrial IoT networks
Federated Learning (FL) empowers Industrial Internet of Things (IIoT) with distributed
intelligence of industrial automation thanks to its capability of distributed machine learning …
intelligence of industrial automation thanks to its capability of distributed machine learning …
Scalable graph convolutional network training on distributed-memory systems
Graph Convolutional Networks (GCNs) are extensively utilized for deep learning on graphs.
The large data sizes of graphs and their vertex features make scalable training algorithms …
The large data sizes of graphs and their vertex features make scalable training algorithms …
Dynamic layer-wise sparsification for distributed deep learning
Distributed stochastic gradient descent (SGD) algorithms are becoming popular in speeding
up deep learning model training by employing multiple computational devices (named …
up deep learning model training by employing multiple computational devices (named …
D3-GNN: Dynamic Distributed Dataflow for Streaming Graph Neural Networks
R Guliyev, A Haldar, H Ferhatosmanoglu - arXiv preprint arXiv:2409.09079, 2024 - arxiv.org
Graph Neural Network (GNN) models on streaming graphs entail algorithmic challenges to
continuously capture its dynamic state, as well as systems challenges to optimize latency …
continuously capture its dynamic state, as well as systems challenges to optimize latency …
A lightweight self-supervised representation learning algorithm for scene classification in spaceborne SAR and optical images
X Xiao, C Li, Y Lei - Remote Sensing, 2022 - mdpi.com
Despite the increasing amount of spaceborne synthetic aperture radar (SAR) images and
optical images, only a few annotated data can be used directly for scene classification tasks …
optical images, only a few annotated data can be used directly for scene classification tasks …
Self-Compressing Neural Networks
S Cséfalvay, J Imber - arXiv preprint arXiv:2301.13142, 2023 - arxiv.org
This work focuses on reducing neural network size, which is a major driver of neural network
execution time, power consumption, bandwidth, and memory footprint. A key challenge is to …
execution time, power consumption, bandwidth, and memory footprint. A key challenge is to …
Mapping and optimization method of SpMV on Multi-DSP accelerator
S Liu, Y Cao, S Sun - Electronics, 2022 - mdpi.com
Sparse matrix-vector multiplication (SpMV) solves the product of a sparse matrix and dense
vector, and the sparseness of a sparse matrix is often more than 90%. Usually, the sparse …
vector, and the sparseness of a sparse matrix is often more than 90%. Usually, the sparse …
SpComm3D: A Framework for Enabling Sparse Communication in 3D Sparse Kernels
N Abubaker, T Hoefler - arXiv preprint arXiv:2404.19638, 2024 - arxiv.org
Existing 3D algorithms for distributed-memory sparse kernels suffer from limited scalability
due to reliance on bulk sparsity-agnostic communication. While easier to use, sparsity …
due to reliance on bulk sparsity-agnostic communication. While easier to use, sparsity …
FSD-Inference: Fully Serverless Distributed Inference with Scalable Cloud Communication
J Oakley, H Ferhatosmanoglu - arXiv preprint arXiv:2403.15195, 2024 - arxiv.org
Serverless computing offers attractive scalability, elasticity and cost-effectiveness. However,
constraints on memory, CPU and function runtime have hindered its adoption for data …
constraints on memory, CPU and function runtime have hindered its adoption for data …
Leveraging Memory Copy Overlap for Efficient Sparse Matrix-Vector Multiplication on GPUs
G Zeng, Y Zou - Electronics, 2023 - mdpi.com
Sparse matrix-vector multiplication (SpMV) is central to many scientific, engineering, and
other applications, including machine learning. Compressed Sparse Row (CSR) is a widely …
other applications, including machine learning. Compressed Sparse Row (CSR) is a widely …