TOMDS (Topic-Oriented Multi-Document Summarization): Enabling Personalized Customization of Multi-Document Summaries

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The document under consideration requires substantial revisions before it can be considered suitable for publication. Several areas have been identified that require attention and improvement, including but not limited to:

1) The results and discussion section lacks sufficient significance and needs significant revisions.

2) The introduction section needs to be improved.

3) Trivial data can be transferred to supporting documents to streamline the main document.

4) Future plans should be included to provide a more comprehensive view of the study.

 

5) The abstract and conclusion section should contain the main numerical results to make the study's findings more apparent.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This paper proposes the topic-oriented multi-document summarization model. The model contains two stages, one is the extractive stage and the other is generation state. The experimental results showed that the proposed model outperforms all the baseline models.

 

This paper is well organized and deals with interesting topic. Especially, the proposed model is described enough to understand details. However, I have several concerns especially about related works and experiments.

It would be much better if the authors address the following comments.

 

Major revision

- [Abstract]

- In the argument "existing technologies cannot consider topic consistency", I don't understand the mean of "topic consistency mean". That is not described in the anywhere of the paper.

 

[Section 1]

- It would be good if the authors describe contributions of this paper in the Introduction section.

- At the final sentence, it would be good if the authors include quantitative experimental results.

- At the final sentence, the authors use the word, "ensure". I think this word is not appropriate in this case because experiments the authors conducted seem not to ensure anything.

 

[Section 2]

- There are a lot of recent studies relevant to topic based multi-document summarization methods. But this paper does not review them except only one paper (Liu et al. [15]) even though they are highly relevant to this paper. The authors must review those papers in-depth.

- My another big concern is that the Related Work section does not provide any useful insights. Sentences just list studies and few descriptions. The authors should describe the limitations of the studies or discuss the relevant and differences with the model the authors propose.

 

[Section 3]

- In Section 3.1, I don’t understand “primary and secondary relationships between paragraphs.” It would be good if an example is provided.

 

[Section 4]

- The proposed model is compared with six models, which are relevant to the proposed one. But, they are not handled in the Related work section.

- The proposed model is topic-based multi-document summarization model. However, it is not compared with topic-based ones. Is there any reason for that?

- In Section 4.4., in the sentence, “more accurately generates An abstract consistent with the original text”, I don’t understand the mean of “consistent with the original text.” Semantically the same? or grammatically correct?

- In Section 4.3, Who participated in the experiment and how were they selected? Are they people who can produce unbiased results? Also, for better understanding, it would be good if sample summarization results for each model are provided and discussed.

 

Minor revision

- In Abstract, maybe "cannot" -> "do not"

- In Section 3.1, abbreviation “EDU” is used without description.

- In Section 3.3.3, Discourse Grapg Attention à Discourse Graph Attention

- In Section 4.1.3, incomplete sentence “Compare it with the paragraph to calculate the ROUGE-2recall rate, and use the logistic regression model LSTM for prediction.”

- In Section 4.1.4, I recommend models to be compared are described with citations

- In Section 4.4., change an uppercase letter to lowercase letter in the sentence “more accurately generates An abstract consistent with the original text”

Comments on the Quality of English Language

I recommend the authors use English proof reading service.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Attached ,  you  can  find  my  comments !

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Moderate  editing  English  required !

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

1. I commend the authors for addressing the significant challenge of topic consistency in multi-document summarization. Their innovative approach leveraging a topic-oriented multi-document summarization (TOMDS) model, incorporating both extractive and abstractive stages with discourse structure analysis, represents a valuable contribution to the field. However, there are several areas where the manuscript could be improved:

2.  

3. 1- The paper would benefit from a thorough proofreading to correct typographical errors (e.g., "Raleted Work" should be "Related Work") and improve readability. Some sections are dense and could be made more accessible with additional explanations or examples.

4.  

5. 2- While the TOMDS model is intriguing, the manuscript could provide more detailed explanations of the model's components, especially the discourse parsing module and the topic attention mechanism. Clarifications on the implementation details and parameter settings would enhance reproducibility.

6. Experimental Validation: The experimental results demonstrate the model's potential, but a deeper analysis of the model's performance in different scenarios or with varied dataset characteristics would be insightful. Additionally, including case studies or qualitative examples of the summaries generated could illustrate the model's practical effectiveness.

7.  

8. 3-The comparison with other models is valuable, yet further discussion on why the TOMDS model outperforms others in specific aspects would be beneficial. Insights into the model's limitations and potential areas for improvement would also be appreciated.

9.  

10. 4- Expanding on the practical applications of the TOMDS model and its impact on real-world summarization tasks, especially in the context of personalized information retrieval, would highlight the significance of this work.

Comments on the Quality of English Language

The English language in the document is generally clear and conveys the intended scientific concepts effectively. However, there are minor typographical errors as mentioned in pervious comments and instances where the language could be refined for better clarity and readability. Attention to detail in proofreading and possibly simplifying complex sentences could enhance the overall presentation and accessibility of the content for a broader audience.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have satisfactorily and thoroughly addressed all my comments from the previous round of reviews.

I do not have any further observations to make. I consider that the manuscript is in good shape for publication.

Reviewer 4 Report

Comments and Suggestions for Authors

Authors have addressed all concerns successfully.