Name: Ahmed M. A. Sayed
Type: User
Company: Queen Mary University of London
Bio: aka. Ahmed M. Abdelmoniem - Assistant Professor at QMUL, UK - Head of SAYED Systems Group - Interested in ML, networking and Distributed Systems
Twitter: ahmedcs982
Location: London
Blog: www.eecs.qmul.ac.uk/~ahmed
Ahmed M. A. Sayed's Projects
Implementation of Parameter Server using PyTorch communication lib
Virtualized Receive Window based DCTCP
A tutorial on RDMA based programming using code examples
Resource Efficient Federated Learning
[ICLR 2020] Contrastive Representation Distillation (CRD), and benchmark of recent knowledge distillation methods
RWNDQ is an Equal Share Allocation Switch Design for Data Centre Networks
SDN-based Transport-Agnostic Congestion Control (SDN-GCC)
SDN proactive fault handling
Source Code for ICML 2019 Paper "Shallow-Deep Networks: Understanding and Mitigating Network Overthinking"
PyTorch library to facilitate development and standardized evaluation of neural network pruning methods.
SDN-based Incast Congestion Control for Data Centers
SIDCo is An Efficient Statistical-based Gradient Compression Technique for Distributed Training Systems
Code for the signSGD paper
Simplicial-FL to manage client device heterogeneity in Federated Learning
Sketched SGD
Accelerating Distributed Machine Learning with Data Sketches
This repository includes supervised and unsupervised machine learning methods which are used to detect anomalies on network datasets. Decision Tree, Random Forest, Gradient Boost Tree, Naive Bayes, and Logistic Regression were used for supervised learning. K-Means was used for unsupervised learning.
Code for Sparsified SGD.
SPATL: Salient Prameter Aggregation and Transfer Learning for Heterogeneous Federated Learning
Investigating Split Learning and Federate Learning