HKU Data Repository
Browse

plot-*.py and 'data' folder

Reason: Since one paper included in the thesis is currently under review, I would like to apply an embargo period to protect data privacy for 2 years.

2

month(s)

9

day(s)

until file(s) become available

Supporting data for “Optimizing Federated Learning with Communication Reduction and Synchronization Control”

dataset
posted on 2023-03-13, 01:29 authored by Xueyu Wu

Experimental data used in Thesis  “Optimizing Federated Learning with Communication Reduction and Synchronization Control”.

There are source codes that read the logs from the 'data/' 

folder and then display the figures or tables.


This thesis aims to improve the federated learning systems with communication reduction techniques and accelerate model convergence speed by asynchronous training. 

The provided data show the effectiveness of our proposed algorithm, which includes two fast K asynchronous algorithms and a novel weighted aggregation function.




History

Usage metrics

    Research Postgraduates

    Exports

    RefWorks
    BibTeX
    Ref. manager
    Endnote
    DataCite
    NLM
    DC