A Cross-Modal Dynamic Attention Neural Architecture to Detect Anomalies in Data Streams from Smart Communication Environments

Round 1

Reviewer 1 Report

The topic of the article paper is timely and interesting. The proposed method is also interesting and involves a combination of technology from recent learning systems. However, the way the authors go about it does not reflect scientific writing and structure. I recommend that authors put more effort into improving the style of their article. Some Important Comments.

1) The intro part needs some kind of justification. The clarifications in the introductory section are so serious and some references should be added to provide strong proof of the claims.

2) Literature review references are relevant to the topic. However, further analysis of these references is needed. Because the mission here is to extract the gaps in the research (i.e. what the authors did not do and should be done). Please list research gaps accordingly.

3) The contributions of this paper should also be listed and thoroughly analyzed and directed towards addressing research gaps.

4) The figures in this article are treated with low quality. I recommend providing artistic and colorful figures in this case with better details and resolutions.

5) "Materials" meaning the case study in this case is provided with such a light description. I recommend a detailed description and illustrations also of the data variation and its scutters!! This will be of great benefit illustrating the complexity of the problem solved. Also, it is important if you show some examples of normal data and data with anomalies!!

6) The application results in Table 1 look acceptable, but since there is no other illustration like data scutters, confusion matrices, loss curves, there is no strong evidence of real performances of the proposed method. I recommend adding such details.

Author Response

Dear respected Reviewer,

We deeply appreciate your time and effort in reviewing our manuscript. Your comments are very helpful for revising and improving our paper. We have revised the manuscript considering all the insightful comments to enhance the paper's readability. We believe these changes have strengthened the rationale and importance of our study.

 

The topic of the article paper is timely and interesting. The proposed method is also interesting and involves a combination of technology from recent learning systems. However, the way the authors go about it does not reflect scientific writing and structure. I recommend that authors put more effort into improving the style of their article. Some Important Comments.

1) The intro part needs some kind of justification. The clarifications in the introductory section are so serious and some references should be added to provide strong proof of the claims.

Thank you for this comment. We have incorporated your suggestions and added appropriate references to provide strong evidence for our claims.

2) Literature review references are relevant to the topic. However, further analysis of these references is needed. Because the mission here is to extract the gaps in the research (i.e. what the authors did not do and should be done). Please list research gaps accordingly.

Thank you for your suggestions. We have thoroughly explained the literature review and listed the research gaps based on your suggestions.

3) The contributions of this paper should also be listed and thoroughly analyzed and directed towards addressing research gaps.

We have provided a comprehensive explanation of our proposed approach based on identified research gaps. Thank you for your feedback.

4) The figures in this article are treated with low quality. I recommend providing artistic and colorful figures in this case with better details and resolutions.

Based on your suggestions, we have made changes to the figures by rearranging, enlarging, or altering them. Thank you for your helpful feedback.

5) "Materials" meaning the case study in this case is provided with such a light description. I recommend a detailed description and illustrations also of the data variation and its scutters!! This will be of great benefit illustrating the complexity of the problem solved. Also, it is important if you show some examples of normal data and data with anomalies!!

Thank you for your insightful comment. In the Materials and Methods section, we provide the conceptual proposed architecture that leverages the advantages of cross-modal learning and dynamic attention mechanisms to effectively analyze heterogeneous data streams from different cyber modalities and identify anomalous patterns in real time. We have included in the section 5.2. Dataset a table displaying various anomalies injected in the synthetic dataset. These anomalies represent a broad spectrum of potential attacks and unusual behaviors that the CM-DANA method aims to detect, illustrating the complexity of the problem solved. Also, we strongly believe that we improved the revised paper by adding explanations of the results and visualization plots to make them more understandable and highlight the novelty of the proposed approach.

6) The application results in Table 1 look acceptable, but since there is no other illustration like data scutters, confusion matrices, loss curves, there is no strong evidence of real performances of the proposed method. I recommend adding such details.

Thank you for your helpful feedback. We have added several plots to visualize the results, making them more understandable.

Reviewer 2 Report

The paper is well-organized and contributing. The methodology is presented like a usual way in this research field. However, some of minor improvements may be required:

1) The source of data and their characteristics, including data profiling should be more clearly stated.

2) The visualization of the results may demonstrate the reader to be more understandable.

3) The significant testing of the results should be conducted and shown the strength of the novelty.

Author Response

Dear respected Reviewer,

We deeply appreciate your time and effort in reviewing our manuscript. Your comments are very helpful for revising and improving our paper. We have revised the manuscript considering all the insightful comments to enhance the paper's readability. We believe these changes have strengthened the rationale and importance of our study.

 

The paper is well-organized and contributing. The methodology is presented like a usual way in this research field. However, some of minor improvements may be required:

1) The source of data and their characteristics, including data profiling should be more clearly stated.

We have included a table displaying various anomalies injected in the synthetic dataset. These anomalies represent a broad spectrum of potential attacks and unusual behaviors that the CM-DANA method aims to detect. Thank you for your insightful comment.

2) The visualization of the results may demonstrate the reader to be more understandable.

Thank you for your helpful feedback. We have added several plots to visualize the results, making them more understandable.

3) The significant testing of the results should be conducted and shown the strength of the novelty.

Thank you for your helpful comment. We strongly believe that we improved the revised paper by adding explanations of the results and visualization plots to make them more understandable and highlight the novelty of the proposed approach.

Reviewer 3 Report

The paper, at its core, aims to present an important and relevant model called the Cross-Modal Dynamic Attention Neural Architecture (CM-DANA) for detecting anomalies in smart communication environments. This is indeed an area that demands innovative approaches and the paper's theme is valuable to the scientific community. However, several issues must be addressed to enhance the clarity, depth, and academic rigor of the work.

 

1. Format and Style:

 

The authors should revise the paper to follow a more conventional academic format. Currently, the paper reads like a summary or a list, rather than a cohesive academic discourse. The text appears to exhibit a style akin to what might be produced by a generative language model, offering a summary-like presentation rather than an in-depth, scholarly examination. This may impair readers' understanding and appreciation of the authors' contributions.

 

2. Figure Issues:

 

Figure 2, as mentioned, appears to have been sourced from the internet, possibly leading to copyright issues. The authors must either obtain proper permissions or replace this figure with an original illustration that accurately represents the relevant content.

 

3. Lack of Depth:

 

The paper seems to provide a very general discussion without delving into specific details or demonstrating unique contributions. Most sections appear superficial, particularly the Limitations section in Section 6. A revision should aim to deepen the analysis, focusing on detailed examinations and critical insights.

 

4. Limitations Section:

 

Section 6 requires significant reworking. The current content mainly discusses general limitations of deep learning models rather than the specific limitations of the proposed CM-DANA model. The authors should elucidate what the shortcomings of their proposed model are, providing a balanced view and guiding future research in this direction.

 

5. Content of Section 3:

 

Section 3 is another critical area that needs attention. Currently, it discusses CNNs and Attention in general terms and lacks specific insights or unique contributions from the authors. This section should be revised to clearly highlight the novel aspects of the CM-DANA architecture and how it incorporates CNNs for feature extraction. The general descriptions of CNNs, transformers, and attention mechanisms could be condensed, and more emphasis should be placed on the unique aspects of the authors' methodology.

 

6. Clarity in Explanation:

 

The authors should work on integrating them more cohesively into the overall narrative of the paper. Please Explain how the steps of Section 3 specifically relate to the proposed CM-DANA architecture.

Author Response

Dear respected Reviewer,

We deeply appreciate your time and effort in reviewing our manuscript. Your comments are very helpful for revising and improving our paper. We have revised the manuscript considering all the insightful comments to enhance the paper's readability. We believe these changes have strengthened the rationale and importance of our study.

 

The paper, at its core, aims to present an important and relevant model called the Cross-Modal Dynamic Attention Neural Architecture (CM-DANA) for detecting anomalies in smart communication environments. This is indeed an area that demands innovative approaches and the paper's theme is valuable to the scientific community. However, several issues must be addressed to enhance the clarity, depth, and academic rigor of the work.

1. Format and Style:

The authors should revise the paper to follow a more conventional academic format. Currently, the paper reads like a summary or a list, rather than a cohesive academic discourse. The text appears to exhibit a style akin to what might be produced by a generative language model, offering a summary-like presentation rather than an in-depth, scholarly examination. This may impair readers' understanding and appreciation of the authors' contributions.

Thank you for your insightful feedback. We strongly believe that we have improved the revised paper by adding appropriate explanations to make it more understandable and highlighting the novelty of the proposed approach. We have also followed a more conventional academic format based on your suggestions.

2. Figure Issues:

Figure 2, as mentioned, appears to have been sourced from the internet, possibly leading to copyright issues. The authors must either obtain proper permissions or replace this figure with an original illustration that accurately represents the relevant content.

Thank you for your helpful comment. This figure is licensed under the Creative Commons Attribution-ShareAlike 3.0 (CC BY-SA 3.0) license. This means that it is free to use, share, and adapt the figure for any purpose, even commercially, as long as proper attribution is provided to the original creator and any derivative works are distributed under the same license terms. This license promotes collaboration, sharing, and the creation of new works while ensuring that the original creator receives credit and that the resulting works remain open and accessible to others.

3. Lack of Depth:

The paper seems to provide a very general discussion without delving into specific details or demonstrating unique contributions. Most sections appear superficial, particularly the Limitations section in Section 6. A revision should aim to deepen the analysis, focusing on detailed examinations and critical insights.

Thank you for your insightful comment. In the revised paper we have deepened the analysis, focusing on detailed examinations and critical insights. Specifically, in the introduction we have added appropriate references to provide strong evidence and proofs for our claims. We have thoroughly explained the literature review and listed the research gaps based on the state of the art. We have provided a comprehensive explanation of our contributions based on identified research gaps. Also, in the Materials and Methods section, we provide the conceptual proposed architecture that leverages the advantages of cross-modal learning and dynamic attention mechanisms to effectively analyze heterogeneous data streams from different cyber modalities and identify anomalous patterns in real-time. In addition, section 3 has been reorganized to emphasize the distinctive and innovative features of the proposed methodology. Moreover, we have included in section 5.2. Dataset, a table displaying various anomalies injected in the synthetic dataset. These anomalies represent a broad spectrum of potential attacks and unusual behaviors that the CM-DANA method aims to detect, illustrating the complexity of the problem solved. Also, we have added several plots to visualize the results, making them more understandable. Finally, we have reorganized section 6 to highlight the limitations of their proposed model, providing a balanced view and guiding future research.

4. Limitations Section:

Section 6 requires significant reworking. The current content mainly discusses general limitations of deep learning models rather than the specific limitations of the proposed CM-DANA model. The authors should elucidate what the shortcomings of their proposed model are, providing a balanced view and guiding future research in this direction.

We have reorganized section 6 to highlight the limitations of their proposed model, providing a balanced view and guiding future research. Thank you for this comment.

5. Content of Section 3:

Section 3 is another critical area that needs attention. Currently, it discusses CNNs and Attention in general terms and lacks specific insights or unique contributions from the authors. This section should be revised to clearly highlight the novel aspects of the CM-DANA architecture and how it incorporates CNNs for feature extraction. The general descriptions of CNNs, transformers, and attention mechanisms could be condensed, and more emphasis should be placed on the unique aspects of the authors' methodology.

Thank you for your helpful feedback. Section 3 has been reorganized to emphasize the distinctive and innovative features of the proposed methodology.

6. Clarity in Explanation:

 The authors should work on integrating them more cohesively into the overall narrative of the paper. Please Explain how the steps of Section 3 specifically relate to the proposed CM-DANA architecture.

Thank you for your helpful comment. The four modules and how related to the proposed CM-DANA methodology are described in section 3. These modules introduce several innovative aspects that collectively enhance the accuracy and efficiency of the proposed architecture, as explained in the revised paper.

Round 2

Reviewer 1 Report

The reviewers almost responded to my comments appropriately. the current version of the paper seems better. I have no additional comments.

Author Response

Dear respected Reviewer,

We deeply appreciate your time and effort in reviewing our manuscript.

Reviewer 3 Report

In the revised manuscript, the authors have addressed the concerns I previously raised in a satisfactory manner. Nevertheless, residual issues still persist, particularly in regard to the stylistic approach employed within the paper. Upon closer inspection, it is apparent that the writing diverges from traditional academic conventions and bears resemblance to a format more commonly found in blog posts or technical reports. This divergence undermines the scholarly tone and potentially the credibility of the research. I would therefore strongly recommend a comprehensive revision to align the manuscript with the established norms of academic writing.

 

I must direct the authors' attention to specific lines in the manuscript that require rectification:

 

Lines 336 - 356, 406 – 422: Using unnecessary list-based writing methods.

Lines 539 - 562: These sections contain content that is too general to be appropriately addressed in scholarly papers. Moreover, the content's relevance to the overall methodology of the paper is tenuous at best.

Lines 572 - 599: The narrative in this section suffers from the same issue, with a general focus that does not befit an academic document, and a lack of clear connection to the paper's methodology.

Lines 613 - 647, 823 - 846: These sections necessitate a comprehensive revision to align with accepted standards of academic writing.

 

It is of vital importance to note that the manuscript frequently adopts a list- based writing structure. This format is incongruent with conventional academic writing methodologies and could diminish the overall credibility of the work. Consequently, I strongly recommend a meticulous review of the aforementioned areas and a reassessment of the paper's overall stylistic approach to ensure alignment with scholarly writing norms.

Author Response

Dear respected Reviewer,

We deeply appreciate your time and effort in reviewing our manuscript. Your comments are very helpful for revising and improving our paper. We have revised the manuscript considering all the insightful comments to enhance the paper's readability. We believe these changes have strengthened the rationale and importance of our study.

 

In the revised manuscript, the authors have addressed the concerns I previously raised in a satisfactory manner. Nevertheless, residual issues still persist, particularly in regard to the stylistic approach employed within the paper. Upon closer inspection, it is apparent that the writing diverges from traditional academic conventions and bears resemblance to a format more commonly found in blog posts or technical reports. This divergence undermines the scholarly tone and potentially the credibility of the research. I would therefore strongly recommend a comprehensive revision to align the manuscript with the established norms of academic writing.

I must direct the authors' attention to specific lines in the manuscript that require rectification:

Lines 336 - 356, 406 – 422: Using unnecessary list-based writing methods.

Thank you for your feedback. We have revised the manuscript based on your suggestions in order to align with academic writing norms.

Lines 539 - 562: These sections contain content that is too general to be appropriately addressed in scholarly papers. Moreover, the content's relevance to the overall methodology of the paper is tenuous at best.

Thank you for providing your knowledge. We have revised the manuscript and removed the specific content that was mentioned.

Lines 572 - 599: The narrative in this section suffers from the same issue, with a general focus that does not befit an academic document, and a lack of clear connection to the paper's methodology.

Thank you for your feedback. We have revised the manuscript based on your suggestions in order to align with academic writing norms.

Lines 613 - 647, 823 - 846: These sections necessitate a comprehensive revision to align with accepted standards of academic writing.

Thank you for your feedback. We have revised the manuscript based on your suggestions in order to align with academic writing norms.

It is of vital importance to note that the manuscript frequently adopts a list- based writing structure. This format is incongruent with conventional academic writing methodologies and could diminish the overall credibility of the work. Consequently, I strongly recommend a meticulous review of the aforementioned areas and a reassessment of the paper's overall stylistic approach to ensure alignment with scholarly writing norms.

We have restructured the paper according to the reviewer's feedback, improving the readability and elevating the work's scientific and academic standing for this journal's readership. Thank you for your careful consideration.