An Improved Chinese Pause Fillers Prediction Module Based on RoBERTa

Round 1

Reviewer 1 Report

Comments for author File: Comments.pdf

Moderate editing of the English language required

Author Response

Dear reviewer,

Thank you very much for your comments and professional advices. These opinions help to improve academic rigor of our article. Based on your suggestion and request, we have made corrected modifications on the revised manuscript. Meanwhile, the manuscript had be reviewed and has undergone English language editing by MDPI. We hope that our work can be improved again.

Furthermore, we would like to show the details in the attachment. We are uploading a Cover Letter for reviewer1.docx (response to reviewers).

Thank you very much for your attention and time. Look forward to hearing from you.

Best regards,

Xiaoqun ZHOU, Ling YU, and Fanglin NIU.

Author Response File: Author Response.docx

Reviewer 2 Report

This is a novel study using RoBERTa model to predict Chinese pause fillers and introduced new approach to train the RoBERTa model.

These are some comments to improve the article.

Page 1, line 27. Pause fillers, also known as pause fillers, are brief pauses or meaningless… Pause fillers – also know as pause fillers…. What do you mean? Same spelling.

Page 1, line 44. In recent years, with the advancement of deep learning technologies, neural network models such as LSTM, BERT, and Transformer have been… different from the abstract in line 13 - neural networks such as LSTM, BERT, and XLNet have been….

Page 2, line 94. Acronym of TTS was introduced in line 92. The the full Text-to-Speech word was used in line 94, line 95, line 101, line 107, line 109, line 124 and a lot more after that. Then what’s the point of introducing the acronym TTS? Before that, the Text-to-Speech was used 5 times in line 74, 77, 78, 81, and 87.

Page 4, Table 1. Row 5- this: Turn down, that, the brightness of this phone a bit. Example is not the same as the pause filler. Row 7 – oho: And playing games, uh, chatting, uh, and browsing videos. Example is not the same as the pause filler. Row 9 – oh: This indicates, uh, that their laws are not as good as ours. Example is not the same as the pause filler.

Page 5, line 184 to 194. The paragraph is exactly the same as Page 4, line 156 to 165. Redundant.

Page 5, line 215. The number 4 is continuation from number 3 above but exactly the same word with number 1. Number 5 is exactly the same as number 3. Should start with a new numbering if separate thing.

Page 6, line 233. Why the number 8? If the continuation, the number 8, 9, and 10 are exactly the same to number 4, 5, and 6.

Page 6, line 238. Figure 1 was not mentioned in the text.

Page 6, line 239. Fig. 2 but page 10, line 352 – Figure 4. Please standardise, use Fig. or Figure?

Page 7, line 248. Fig. 3 should be Fig. 2 based on the figure caption.

Use past tense instead of present tense.

Author Response

Dear reviewer,

Thank you very much for your comments and professional advices. These opinions help to improve academic rigor of our article. Based on your suggestion and request, we have made corrected modifications on the revised manuscript. Meanwhile, the manuscript had be reviewed and has undergone English language editing by MDPI. We hope that our work can be improved again.

Furthermore, we would like to show the details in the attachment. We are uploading a Cover Letter for reviewer2.docx (response to reviewers).

Thank you very much for your attention and time. Look forward to hearing from you.

Best regards,

Xiaoqun ZHOU, Ling YU, and Fanglin NIU.

Author Response File: Author Response.docx

Reviewer 3 Report

1. There are many repetitions between abstract and conclusion. Therefore, this abstract part needs to be rewritten.

2. There are many repetitions between abstract and conclusion. Therefore, this conclusion part needs to be rewritten.

3. All the figures in the paper have different fonts, please check carefully and improve them.

4. Discussion is an important part of the article. It is suggested that the authors supplement this section separately.

5. The format of references is not uniform. For example, the pages of some references are missed, the Journal name of some references are wrong, the author name are wrong, and so on.

6. The language of the article needs to be further improved. It is suggested to polish the article before submitting it for publication.

Comments for author File: Comments.docx

The language of the article needs to be further improved. It is suggested to polish the article before submitting it for publication.

Author Response

Dear reviewer,

Thank you very much for your comments and professional advices. These opinions help to improve academic rigor of our article. Based on your suggestion and request, we have made corrected modifications on the revised manuscript. Meanwhile, the manuscript had be reviewed and has undergone English language editing by MDPI. We hope that our work can be improved again.

Furthermore, we would like to show the details in the attachment. We are uploading a Cover Letter for reviewer3.docx (response to reviewers).

Thank you very much for your attention and time. Look forward to hearing from you.

Best regards,

 Ling YU,Xiaoqun ZHOU, and Fanglin NIU.

Author Response File: Author Response.docx

Reviewer 4 Report

The paper is a well-written paper on predicting Chinese pause fillers with the RoBERTa model. Predicting pause fillers was underestimated for a long time. It shows significant improvements compared to other transformers. It would be nice to apply the results also to other languages. Presentation, structure and style is ok. Only a paragraph dedicated to a hypothesis (improved performance, ...) and research goals is missing.  

Only minor grammatical issues observed.

Author Response

Dear reviewer,

Thank you very much for your comments and professional advices. These opinions help to improve academic rigor of our article. Based on your suggestion and request, we have made corrected modifications on the revised manuscript. Meanwhile, the manuscript had be reviewed and has undergone English language editing by MDPI. We hope that our work can be improved again.

Furthermore, we would like to show the details in the attachment. We are uploading a Cover Letter for reviewer4.docx (response to reviewers).

Thank you very much for your attention and time. Look forward to hearing from you.

Best regards,

Ling YU, Xiaoqun ZHOU, and Fanglin NIU.

Author Response File: Author Response.docx

Reviewer 5 Report

The article is well presented, and the context is of interest. However, I have the following concerns.

-          The introduction section is too long; reconsider the section to be more concise. Reconsider adding a new section on background and related works. You may use the data in the introduction section (move a part of it to the new section).

-          All acronyms and initialisms should be defined the first time they appear in the text. This should be in the abstract as well as the article.

-          Consider adding two more relevant keywords.

-          Consider highlighting the main contributions of the article at the end of the introduction section.

-          “The dataset is divided into training, validation, and testing sets in approximately an 8:1:1 ratio.” Does this ratio fit well? Consider adding a citation.

-          Consider defining the performance metrics ( Precision,…).

The article requires extensive proofreading.

Author Response

Dear reviewer,

Thank you very much for your comments and professional advices. These opinions help to improve academic rigor of our article. Based on your suggestion and request, we have made corrected modifications on the revised manuscript. Meanwhile, the manuscript had be reviewed and has undergone English language editing by MDPI. We hope that our work can be improved again.

Furthermore, we would like to show the details in the attachment. We are uploading a Cover Letter for reviewer5.docx (response to reviewers).

Thank you very much for your attention and time. Look forward to hearing from you.

Best regards,

Ling YU, Xiaoqun ZHOU, and Fanglin NIU.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

After modification, the manuscript generally meets the basic requirements of paper writing.

Reviewer 5 Report

All comments have been addressed. I have no more comments, the article is recommended for publication in its current form.

The language style is fine.