A Carrying Method for 5G Network Slicing in Smart Grid Communication Services Based on Neural Network
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe topic of the paper is not fully disclosed, in fact, there is no clear understanding of what the proposed method is. The numerical analysis is mainly devoted to the problem of the accuracy of the application of the neural network, but there is not a single result showing what the proposed method gives for the slicing technology itself. For this, a system model is needed, from which the parametrization and scope of the proposed method would be clear. There is no comparison with other works and methods, although slicing has already become a well-researched technology.
There are also specific critical remarks that do not allow a positive assessment of the study.
1. The activation function of the proposed LSTM and its parameters are not defined.
2. The authors do not present an important result of the distribution of the common channel to services based on the obtained results of the neural network model.
3. There is no evaluation of the performance of the neural network model based on the traffic load of various services. This is the result of the complete absence of a systemic model.
3. Based on the proposed data in the paper, it is recommend comparing the accuracy with the well-known Short-time Fourier transform (STFT) method. Especially for discrete input values (bits), the STFT method will be less computationally intensive. It is also recommended to submit the spectrum from the STFT as input data to the neural network and compare it with the current results of the work.
4. The paper does not describe the motivation for using MEC for the proposed model.
5. There is no visual comparison with other works.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsAlthough the topics of the paper are very interesting there are some major improvements needed:
1. The title is not proper: what do you mean by "Power 5G Slicing Service"?
2. The abstract needs to be reformulated (it should contain complete phrases (with verbs) , not a list of features/issues.
3. There are words repeated too many times in the same phrase, too many times. E.g., "intelligence" in rows 66-69, "communication" in rows 119-120 etc.
4. The equations and the text in Chapter 3 should have references (as I presume the authors did not invent them.
5. A list of acronyms at the end of the paper.
6. Some references do not follow IEEE rules (e.g., [18]).
Comments on the Quality of English Language
1. The abstract needs to be reformulated (it should contain complete phrases (with verbs) , not a list of features/issues.
2. There are words repeated too many times in the same phrase, too many times. E.g., "intelligence" in rows 66-69, "communication" in rows 119-120 etc.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsAlthough the idea of the paper is very interesting, the whole paper needs imrovement to be in a publishable form. Dear authors please consider the following remarks:
1) Are Figures 1, 2 produced by the authors? Otherwise please provide references.
2) Lines 163-168: there are no references.
3) Figure 6: It is not clear which dataset represents 128 and 256 bits.
4) Figure 10 needs more analysis.
5) At the end of section 4 the authors must put a Table clearly comparing their approach with published literature. This Table must claerly indicate the superiority of the proposed approach with published literature.
6) The abstract must be rewritten: at many points the reader can not understand what the authors mean.
7) Line 96: What FA means? Please explain.
8) References for equations 1, 2?
9) Poor quality for Figures 6, 7, 9, 10.
10) Please remove lines 249-252 since they do not provide any information for the reader.
11) The reference list must be expanded: about 30-35 references are fine for a journal publication.
12) At Table 3 please specify the length unit.
13) Lack of references at section 3.2.
14) Please explain data types I, II, III, IV, V. It would be better to use a Table for the different data types and scenarios.
15) At Figure 9 what "epoch" means? Please explain.
16) There are no paragraphs in text at all. Please correct.
17) Line 317: Why 8:2 ratio? Please explain.
18) Figure 9: What is the difference between "loss" and "val_loss"? Please explain.
Comments on the Quality of English LanguageThe whole paper needs English language editing. A few parts of the paper (e.g. abstract) is difficult to understand.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for Authors1) There is the initial part of the abstract that is evidently full of grammatical errors and incorrect phrasing: "To adapt to the new situation of 5G power service demand and service types surge, to 11 ensure the continued and stable development of smart grid information and intelligence, proposed 12 a neural network-based power 5G slicing service bearing method. 5G network slicing technology in 13 the realization of power traffic and other traffic isolation at the same time, but also to meet specific 14 traffic needs and service quality assurance." Please improve it.
2) Before delving into the more specific literature of 5G network slicing algorithms and architectures, the authors should briefly illustrate the main tasks that should be guaranteed in order to enable an efficient network slicing. In the specific case of 5G networks, it is well-known that important aspects that should be guaranteed are, for instance, accurate localization of user equipments, channel state estimation, as well as dedicated beamforming strategies. In this respect, I would suggest adding some pointer to the existing literature in this respect. For instance, low-complexity channel estimation strategies for 5G networks have been presented in "Low-complexity downlink channel estimation in mmwave multiple-input single-output systems", IEEE WCL 2021. Improving this initial part would help a potential reader to better focus on all the challenges that should be jointly tackled to provide an effective network slicing strategy.
3) A table summarising the main advantages/disadvantages of the existing 5G network slicing algorithms discussed in the second part of the Introduction would be helpful to better highlight the main novelty of the proposed bearing method and how much it advances state-of-the-art.
4) Performance results reported in Section 4 should be corroborated by a complexity analysis aimed at quantifying what is the computational cost required by the novel proposed network slicing algorithm based on 1D_CNN and BILSTM neural networks.
Comments on the Quality of English LanguageThere are several grammar error and mistakes in the document. Please perform a careful proofreading of the whole document to meet the required standards for a scientific publication.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsAlmost all the comments of the first round of review have been taken into account. In general, the manuscript is in a satisfactory condition and after a minor revision, which is reflected in a thorough check of the text for typos, etc., it can be accepted for publication.
Author Response
We appreciate the guidance of the reviewer. In this round of revisions, we have improved the quality of some data charts and made corresponding revisions to the explanatory text.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe authors have impoved the paper a lot but still a few points must be considered:
1) Comment 3: Figure 6 needs more analysis: it is not clear if Figure 6 represents one line or two lines (what about 128 and 256?). Please explain.
2) Comment 5: Have not been done. The comparison of the proposed approach could be done with references used at the end of the paper.
3) Comment 7: Please add this abbreviation (FA) at the abbreviation list at the end of the paper.
4) Comment 8: Have not been done.
5) Comment 9: Still poor quality for Figures 6, 7, 10.
Comments on the Quality of English LanguagePlease read again your work for English language mistakes.
Author Response
1) Comment 3: Figure 6 needs more analysis: it is not clear if Figure 6 represents one line or two lines (what about 128 and 256?). Please explain.
Response: Thanks for the reviewer’s comment. Figure 6 shows the change in experimental accuracy as the number of iterations increases. This experiment was obtained with the optimal convolutional kernel parameters, that is, the number of two convolutional kernels were 128 and 256.
128 and 256 are the number of convolutional kernels for the two convolutional layers in the CNN structure, respectively.
2) Comment 5: Have not been done. The comparison of the proposed approach could be done with references used at the end of the paper.
Response: Thanks for the reviewer’s comment. At the end of Chapter 4, we include a comparison with the others approaches in Ref 34.
3) Comment 7: Please add this abbreviation (FA) at the abbreviation list at the end of the paper.
Response: As suggested by the reviewer, we have added the abbreviation (FA) at the abbreviation list at the end of the paper.
4) Comment 8: Have not been done.
Response: Thank you for your corrections. We have added reference 31 and 32.
5) Comment 9: Still poor quality for Figures 6, 7, 10.
Response: Thanks for the reviewer’s comment. We've exported the figures to a higher quality. In Figure 6, we deleted the top axis and the right axis to avoid misunderstanding. In Figure 7, we refined the content to mainly show useful information. We split the figure 10 into two pictures (a) and (b) to make it clearer.
Comments on the Quality of English Language
Please read again your work for English language mistakes.
Response: Thanks for the reviewer’s suggestion. We’ve read the paper again and corrected some English language mistakes.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsAuthors correctly addressed all my comments.
Comments on the Quality of English LanguageN/A
Author Response
We appreciate the guidance of the reviewer. In this round of revisions, we have improved the quality of some data charts and made corresponding revisions to the explanatory text.
Author Response File: Author Response.pdf
Round 3
Reviewer 3 Report
Comments and Suggestions for AuthorsThe manuscript has improved a lot compared to the initial submission. A few points for the authors to consider:
1) Comments 5 and 14 have not been done.
2) Figures 11, 12 have poor quality.
3) Please change last section numbering to 5.
Comments on the Quality of English LanguageThe authors must read carefully the paper and correct any English language mistakes.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 4
Reviewer 3 Report
Comments and Suggestions for AuthorsAll the review suggestions have been considered by the authors.
Comments on the Quality of English LanguageEnglish language is now at an acceptable level. Nevertheless, a final carefull read by the authors is required.
