Blockchain Federated Learning for In-Home Health Monitoring
, Samar Othman Alqahtani 3, Wamda Nagmeldin 3, Ashraf Osman Ibrahim 2 and Abdullah Gani 4
Round 1
Reviewer 1 Report
1. Abstract is unnecessarily long. It should provide concise information.
2. First paragraph: Page 2: it should contain some latest references and strongly relevant information
3. Few Paragraphs in the paper should be merged and some long paragraphs can be split into multiples.
4. To broaden the scope of this paper, the authors should refer to some research such as: Federated Learning for Privacy Preservation of Healthcare Data from Smartphone-based Side-Channel Attacks Risk monitoring strategy for the confidentiality of healthcare information
5. It is highly recommended to list all the used symbols in a table with their descriptions to follow the proposed methodology easily.
6. There parameters of the proposed work should be further discussed, why these sets of parameters are used.
7. Few recent references can be added:
A Novel Federated Fog Architecture Embedding Intelligent Formation, Internet of Things intrusion Detection: Centralized, On-Device, or Federated Learning? , Doi:10.4018/978-1-7998-0186-3.ch013
8. It’s common knowledge that the abbreviations should be given full spellings for the first time. However, the authors fail to do this. For example, in the abstract, the abbreviations ML and AES are not given full spellings.
9. Page no.4 contains a summary table of some works without citation within the text. The summary table needs improvement by adding more evidence from recently published papers.
10.Some references do not contain enough information(For example ref. 1, 20 etc.)
Author Response
- Abstract is unnecessarily long. It should provide concise information.
Updated.
- First paragraph: Page 2: it should contain some latest references and strongly relevant information
Updated and cited.
- Few Paragraphs in the paper should be merged, and some long paragraphs can be split into multiples.
Updated.
- To broaden the scope of this paper, the authors should refer to some research such as: Federated Learning for Privacy Preservation of Healthcare Data from Smartphone-based Side-Channel Attacks Risk monitoring strategy for the confidentiality of healthcare information
done.
- It is highly recommended to list all the used symbols in a table with their descriptions to follow the proposed methodology easily.
Done
- There parameters of the proposed work should be further discussed, why these sets of parameters are used.
Question is not clear.
- Few recent references can be added:
A Novel Federated Fog Architecture Embedding Intelligent Formation, Internet of Things intrusion Detection: Centralized, On-Device, or Federated Learning? , Doi:10.4018/978-1-7998-0186-3.ch013
done.
- It’s common knowledge that the abbreviations should be given full spellings for the first time. However, the authors fail to do this. For example, in the abstract, the abbreviations ML and AES are not given full spellings.
Done
- Page no.4 contains a summary table of some works without citation within the text. The summary table needs improvement by adding more evidence from recently published papers.
Improved and updated.
- Some references do not contain enough information (For example ref. 1, 20 etc.)
Updated.
Reviewer 2 Report
Abstract: Average/ need more clarity (still problem statement are not clear to me)
Keywords: Good
Introduction: Average
Literature Review: Average
Data Set: Average
Methodology: Weak/ please verify the protocols of blockchain transaction execution
Caption, Citations & Footnotes: Good
Pictures, graphs & Flowcharts: Average
Results: Weak/required improvement
Conclusion: Average
Future Work: Poor
References: Average
----------- Overall evaluation -----------
This paper presents a blockchain federated learning-enbaled In-Home health monitoring for privacy and security.
Suggestion and Recommendation:
1 (a). In the introduction, the scientific problem of the existed evaluation is missing. It should be elucidated clearly.
1(b). The formatting of the paper most looks like thesis not a research article. Please improve the paper organization.
2 (a). Please improve introduction section. And so, discard extra details which is not related to the proposed work.
2 (b). At the end of the introduction, it is recommended to clearly state the research goal/objectives of this study and what the authors have done to address the identified research problem/section description. Remove extra description which is not relevant to your proposed system.
3 (a). Please improve writing of this paper,
3 (b). The authors should be rechecking the grammatical errors and typos issues in the complete manuscript.
4. The presented diagram should be modified (required block diagram of the proposed work), try to define more. For example, explain sequences of operational execution (control flows), especially in Figures, such as Figure 1, 2, 3, 4, 5, etc.
5. I suggest you explore more open research issues in this domain and add at least 4 open areas that need experts’ futuristic consideration.
6. Reference format must be uniform.
Authors are encouraged to base on recent references about blockchain federated learning-enbaled In-Home health monitoring using artificial intelligence, such machine learning, deep learning, with consortium blockchain, public/private network and DLTs.
7. Please improve contribution section. Till now, the novelty of the work is not highlighted yet.
8. In methodology, only a single algorithm presented to demonstrate the working operation? authors should be clarifying the events of execution (one-by-one) for the whole process? Please elaborate.
9. The experiments should be expanded including more analysis and comparisons with other indexes/baselines and compare your proposed method with newly state-of-the-art methods.
10. The topic is very good and unique but need to improve paper organization.
Author Response
1 (a). In the introduction, the scientific problem of the existed evaluation is missing. It should be elucidated clearly.
With due respect with tried our best in the introduction section to answer the question
“A network of portable medical devices called the Internet of Medical Things (IoMT) gathers actual medical data on a particular patient. These files are large and sensitive, requiring a secure environment. As blockchain technology develops, researchers are focusing on using it to bring cryptography to healthcare applications. The blockchain network has proven to be the most dependable and decentralized system. It offers such intuitive interfaces such as tamper-proofing, immutability, traceability, data storage, confidentiality, and privacy without third party’s participation. Blockchain technology can be used to prevent data from manipulation, maintain data integrity, and provide secure access, distributed computation, and decentralized data storage. As a conclusion, we describe a secure medical data processing and storage system in this paper that employs blockchain to safeguard client data privacy. Concerns about security and privacy have been addressed using federated learning [3]–[6].”
1(b). The formatting of the paper most looks like thesis not a research article. Please improve the paper organization.
Thank you for this important point; we have reformatted the paper to be a research article.
2 (a). Please improve introduction section. And so, discard extra details which is not related to the proposed work.
We have improved and updated the Introduction section.
2 (b). At the end of the introduction, it is recommended to clearly state the research goal/objectives of this study and what the authors have done to address the identified research problem/section description. Remove extra description which is not relevant to your proposed system.
Done.
3 (a). Please improve the writing of this paper,
We have tried our best to improve the paper.
3 (b). The authors should be rechecking the grammatical errors and typos issues in the complete manuscript.
We are thankful to the reviewers for highlighting it. Extensive grammatical reviewing is performed again.
- The presented diagram should be modified (required block diagram of the proposed work), try to define more. For example, explain sequences of operational execution (control flows), especially in Figures, such as Figure 1, 2, 3, 4, 5, etc.
With due respect Figure 2 is our block diagram.
- I suggest you explore more open research issues in this domain and add at least 4 open areas that need experts’ futuristic consideration.
We have included a section named “Open research Issues”.
- Reference format must be uniform.
Authors are encouraged to base on recent references about blockchain federated learning-enbaled In-Home health monitoring using artificial intelligence, such machine learning, deep learning, with consortium blockchain, public/private network and DLTs.
Done.
- Please improve contribution section. Till now, the novelty of the work is not highlighted yet.
Done - In methodology, only a single algorithm presented to demonstrate the working operation? authors should be clarifying the events of execution (one-by-one) for the whole process? Please elaborate.
Tried to improve in the updated version.
- The experiments should be expanded including more analysis and comparisons with other indexes/baselines and compare your proposed method with newly state-of-the-art methods.
Thank you for this comment, the comparison should be done with other studies that used the same dataset to be a fair comparison. Unfortunately, we collected our own data. To the best of our knowledge, we couldn’t find any similar work with the same experimental setup to compare.
- The topic is very good and unique but need to improve paper organization.
Thanks very much for the encouraging comments; we have tried our best to reorganize it.
Reviewer 3 Report
-The novelty and major contribution of the research need properly described.
-There could be a section or text on the cost of implementing the solution proposed in the manuscript.
-Add a table in section II to present the limitations of the works cited and what the authors are contributing?
-I leave some suggestions for related work: -HOsT: Towards a Low-Cost Fog Solution via Smart Objects to Deal with the Heterogeneity of Data in a Residential Environment; -Exploiting offloading in IoT-based microfog: experiments with face recognition and fall detection; -FlexPersonas: flexible design of IoT-based home healthcare systems targeted at the older adults; -A fog-enabled smart home solution for decision-making using smart objects; -NodePM: A Remote Monitoring Alert System for Energy Consumption Using Probabilistic Techniques; -Energy-efficient smart home systems: Infrastructure and decision-making process; -ResiDI: Towards a smarter smart home system for decision- making using wireless sensors and actuators;
-Section IV left the manuscript to be desired. First, it is not possible to replicate the results. Second, it does not validate the proposed solution. Third, there is no statistical analysis. Finally, there is no comparison with any work in the literature to validate the proposal?
-How many clients were used in FL?
-Why don't the authors compare FL with a centralized model?
-The authors could make the proposed solution available on Github?
-It is necessary to describe the behavior of the results and not just describe the graphics?
Author Response
1-The novelty and major contribution of the research need properly described.
Done in the introduction part.
2-There could be a section or text on the cost of implementing the solution proposed in the manuscript.
Because of the chanining market price, we couldn’t provide an accurate cost to address in the study.
3-Add a table in section II to present the limitations of the works cited and what the authors are contributing?
Done in table 1.(require some work)
4-I leave some suggestions for related work: -HOsT: Towards a Low-Cost Fog Solution via Smart Objects to Deal with the Heterogeneity of Data in a Residential Environment; -Exploiting offloading in IoT-based microfog: experiments with face recognition and fall detection; -FlexPersonas: flexible design of IoT-based home healthcare systems targeted at the older adults; -A fog-enabled smart home solution for decision-making using smart objects; -NodePM: A Remote Monitoring Alert System for Energy Consumption Using Probabilistic Techniques; -Energy-efficient smart home systems: Infrastructure and decision-making process; -ResiDI: Towards a smarter smart home system for decision- making using wireless sensors and actuators;
Done, please see check references.
5-Section IV left the manuscript to be desired. First, it is not possible to replicate the results. Second, it does not validate the proposed solution. Third, there is no statistical analysis. Finally, there is no comparison with any work in the literature to validate the proposal?
This is our future plan to validate the proposed solution and statistical analysis as well.
6-How many clients were used in FL?
500.
7-Why don’t the authors compare FL with a centralized model?
The difference between FL and centralized model let us don’t compare between them due to the comparison may be not fair in some criteria.
8-The authors could make the proposed solution available on Github?
Thank you for this suggestion, we will make the proposed solution available on Github soon.
9-It is necessary to describe the ehaviour of the results and not just describe the graphics?
Tried to explain.
Round 2
Reviewer 2 Report
The authors have addressed all my concerns. I believe this version can be accepted. Thank you for your efforts.
Reviewer 3 Report
This reviewer believes that the manuscript can be accepted for publication
