Fed2A: Federated Learning Mechanism in Asynchronous and Adaptive Modes
Round 1
Reviewer 1 Report
Please see the referee report.
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
The authors have proposed Fed2A: Federated Learning mechanism in Asynchronous and Adaptive Modes.
However some queries need to be addressed as follows.
- The motivation behind Fed2A is not understood at all. What is the need? Make the section in Introduction clear about it.
- Comparative analysis between Fed2A and related works should be presented in related works section.
- The evaluation is conducted based on three standard datasets (i.e., 73
FMNIST, CIFAR-10, and GermanTS. Why only these 3 datasets? Is there any specific reason? State the reason why? - Detailing is needed for Fed2A. Current description is hard to follow. Readability should be improved
- You should provide algorithmic complexity analysis.
- More in-depth result aware analysis must be paved.
- Conclusion section is too weak.
- No future scope neither existing challenges mentioned.
- More timely references should be updated.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The submitted Manuscript presents a Federated Learning mechanism in Asynchronous and Adaptive Modes. Their findings of unreported characters are useful for Asynchronous and Adaptive Modes. However, there are a number of minor short-comings there require some attention before final decision.
Author clearly state the detailed introduction and literature to show the major objectives and available technologies in the prescribed area. There is no clear idea how the FL has been applied in various domains.
What is the contribution of this work? It is such a short description with no clue of any contribution to the body of knowledge. No clarity with proposed description.
“To fill the gap, this paper proposes Fed2A”, There is proper evidence for the same. Must clear with your justification.
Give the overall architecture/network diagram to show the work flow of Fed2A deployments.
State the comparison method of all conventional techniques associated with Fed2A Computing.
Mathematical equations not bring the originality of the working procedure of the Fed2A approach. Recommended to match with current problem definition
Many equations are derived, but all shown as general problems, No specific problem addressed.
Suggested separate discussion section to address all the performance measures used in this proposed Fed2A analysis.
The final conclude is very formal. This would benefit from slightly more reflective discussion of each section balancing contributions, shortcomings and potential for further work.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
-General Evaluation: it can be accepted after some Major corrections.
-We request the authors to revise the paper according to the following comments and responses carefully to all the comments to ensure that the quality of this paper has been improved.
-Then we can re-evaluate it for the potential of the next process.
-Check the paper language and make sure all language errors have been fixed. This will help the reader in the future to be more interested in getting the details and knowledge from this paper and resulting in more citations and reputation.
-The abstract section needs significant improvement to be clearer and more attractive for future readers. (rewrite the abstract to reflect the main idea and it's results, without any not suitable details) in the ABSTRACT alongside with the obtained results (the results you got it and what is the situation of your results in comparison with other published methods). ----Mentioned to the benchmarks which have been used in this paper.
-The research findings and contribution need to be stated clearly. As well as, the obtained results in this paper. So, the authors are requested to connect the main idea and contribution with the obtained results to prove that the proposed method accomplished its main aims and better results.
-When referring to related work in the bibliography, adding citations to well-known worldwide journals (especially publications in recent years) would inspire people in its research community to take an interest in this presentation. For example, the following papers might be cited in your work.
https://www.mdpi.com/1424-8220/21/10/3335
https://www.sciencedirect.com/science/article/pii/S0045782520307945
https://ieeexplore.ieee.org/abstract/document/8843942
https://www.sciencedirect.com/science/article/pii/S0360835221001546
https://www.sciencedirect.com/science/article/pii/S0957417421014810
-An overview of the related works is needed, which can show future readers more details about the problems and the standard methods that have been used to solve similar problems. This can support your work and show the reader how you decided to give the new idea based on some avilible information.
The discussion was too shallow and did not explain why the proposed method was superior. The authors are also requested to focus on the obtained results and reflect the proposed method's effect on them. The robustness about the method have not been discussed.
-Check the mathematical notation, especially for the proposed method. This will facilitate the new readers' tracking and applying of the proposed method and get the same results.
-In the conclusion section, it will include research contributions, research limitations, and future works. This part should summarize the whole paper and give the readers full screen to understand the idea, work, contribution, problem, results, new potential work, and others. What are the pros and cons of the proposed method ? Please respond to this question in the article text.
References have not been written in a same format according to the journal reference writing style. They have been listed untidily.
More experiments can be added to boost the paper quality and a comparison with other methods may be needed
What is the difference between your work and this work: " FedSA: A semi-asynchronous federated learning mechanism in heterogeneous edge computing"
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
The revisions are are made.
Reviewer 3 Report
Recommended
Reviewer 4 Report
it can be accepted
