Multi-Class Transfer Learning and Domain Selection for Cross-Subject EEG Classification
Round 1
Reviewer 1 Report
· Spell out each acronym the first time used in the body of the paper. Spell out acronyms in the Abstract.
· What is the motivation of the proposed work? Research gaps, objectives of the proposed work should be clearly justified.
· Introduction needs to explain the main contributions of the work more clearly.
· The novelty of this paper is not clear. The difference between present work and previous Works should be highlighted.
· In the references in the Introduction section, the authors cite some works. However, they have not indicated the advantage or disadvantage and their relations to this paper. It’s a little confusing.
· The author has mentioned the errors obtained by used techniques, it is suggested that the significance of errors listed, must be described in the comparison section.
· The major trends of the simulation should be noted using bullet points.
· Comparsion with recent study and methods would be appreciated.
· Literature review techniques has to be strengthened by including the issues in the current system and how the author proposes to overcome the same.
· Need detailed explanation of the preprocessing steps.
· Clarify the finding Error rate and accuracy in performance analysis section.
· Introduce the chart for given algorithm with description.
· Experimental results are not clear. What are the parameters used in the proposed system and how their values are set? Also, how the parameter values can affect the proposed system? Sections like Experimentation have to be extended and improved thus providing a more convincing contribution to the paper.
· The authors provided details about the implementation setup and working environment. However, some training info should also be given in experimental section. How long does the proposed approach take to learn parameter? These details are missing and must be added to keep the paper standalone.
· an error and statistical analysis of data should be performed.
· a comparison with state of art in the form of table should be added
· more motivation/context regarding the application side of it, particularly on the aspects that make this technique particularly suited for industrial application scenarios, and how it would be applied in real scenarios. These aspects could additionally be supported with some related work context.
Correlate it with other current Technologies, such as: IoT (communications, networks, Cloud, …), in terms of latency I guess that this field is quite sensitive to the delays required to process data, which should call for new investigations around the tradeoff between learning cost and performance (e.g. Deep Learning is costly, yet attains good predictive scores… should we opt for weak learners over good features? Or complex learners over raw data? Or a mixture of both of them, e.g. learned features off-line + weak learners on-line? Should data be sent to the cloud? Be preprocessed at the edge?). This issue is also very trendy at the communications level.
I would also suggest including aspects related to data such as ethics/explainable AI, which are of utmost importance in this domain. Also, the lack of annotated data of admissible quality is a problem for certain problems, for which data augmentation techniques, transfer learning and domain adaptation are for sure fields within AI that the community should aim at in the near future.
Author Response
Author's response to reviewer 1 comments.
Author Response File: Author Response.pdf
Reviewer 2 Report
1)The article is disorganized. The written is more like a course report not a academic article
2)Should compare with more transfer learning baselines, not only traditional machine learning methods
Author Response
Author's response to reviewer 2 comments.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper proposes Multi-Class Transfer Learning and Domain selection for Cross-Subject EEG Classification, which focuses on hot issues, but there are some specific issues in the article as follows:
introduction:
1. The author has cited many references that are too old, and it is recommended to select journals from the past five years;
The introduction to previous work lacks logic and does not well reflect the current problems to be solved. The authors have failed to demonstrate the necessity of their proposed methods through the introduction.
Algorithm part:
The accompanying drawings contain a large amount of text information, and it is recommended to express them in the format of vector diagrams;
The article describes the existing algorithms in detail, with sufficient formulas, but lacks innovation in the algorithm.
Experimental part:
Another obvious problem with this article is the lack of a sufficient explanation of the results. Authors need to explain in detail their experimental results and why you obtained such results.
Author Response
Author's response to reviewer 3 comments.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The article deserves to be accepted in its present form.
Author Response
Author response to reviewer-1
Author Response File: Author Response.pdf
Reviewer 2 Report
I have no further questions
Author Response
Author response to reviewer-2
Author Response File: Author Response.pdf
Reviewer 3 Report
The writing of the manuscript is poor. The experiment results are not convinced,more related and recenct methods should be compared and more experiments with different conditions should be conducted.
Author Response
Author response to reviewer-3
Author Response File: Author Response.pdf