Chinese–Vietnamese Pseudo-Parallel Sentences Extraction Based on Image Information Fusion
Round 1
Reviewer 1 Report
Abstract
The summary lacks precision. Could the authors add a sentence positioning the work from an upstream theoretical point of view? Also, could the authors make explicit the controversy that exists in this regard?
Introduction
Could the authors start the theoretical introduction with a reminder of the definition of parallel sentence pairs ?
Related works
The authors juxtapose previous research on the subject, but there seems to be a lack of perspective on these studies. What are the strengths and weaknesses of each? What are the target languages for each study etc.? I think it would be interesting to develop this section further so that the reader understands what is at stake in this study.
Experiment :
Could the authors say more about the corpus? How were the "10,000 sentence-image pairs to construct the subject image look-up table" selected and what methodological precautions did the authors take?
Conclusion
The conclusion seems really limited. Could the authors make a synthesis of the results obtained, the strong points, the limits of the method. Similarly, could the authors make an explicit link with previous studies on the subject?
It also seems important to propose some elements of perspective applications to this study.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Chinese-Vietnamese Pseudo-parallel Sentences Extraction Based on Image Information FusionIn the considered manuscript, the authors propose a novel method for sentences extraction, to be used in machine translation. The method is aimed at collections of bilingual texts (Chinese-Vietnamese in the particular case) and is based on image information fusion. Generally, I liked the paper, and it has all the necessary components for a solid publication: a clear and practical topic, description of the method and the evaluation. It is also reasonably well written and easy to follow. There are some relatively minor issues that I'd like to highlight, but overall I think it could be accepted to Information.
1) Although most of the needed detail is provided, I feel that the manuscript could be extended to qualify for a journal paper.
2) One of the parts that clearly deserve extending is Discussion. This section does not exist in the current version of the manuscript, so authors do not present the limitations of their study and, arguably more importantly in practice, of applicability of their method. For instance, how would it work with other language pairs? Is it totally universal in this regard?
3) The sources used in the paper are rather dated.
Most importantly, the baseline [17] chosen for the comparison (Table 3) is from 2018 (over 5 years ago, which is a lot of time in ML). The authors should evaluate against a newer baseline to demonstrate significance of their contribution.
The references must be updated, currently just 1 out of 17 is from the last 3 years.
Minor:
Fig. 1: the numbers on the X-axis would be more readable if the text is vertical.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Thank you
Author Response
Thank you for taking the time to provide us with your valuable feedback. If there is anything else we can do to improve the report, please do not hesitate to let us know.
Reviewer 2 Report
I have read the authors' replies to my comments and the comments of the other reviewer. I'm glad to see that our concerns were addressed, and I commend the authors for their great work. Although the paper is still a bit short in my opinion, I believe it can be accepted.
A few minor issues:
* typo: "feed forwOrd" in Fig. 1 (p. 4)
* Fig. 1 that is in p. 9 should be Fig. 2
English is fine, just check for typos (like the ones I noticed in Fig. 1).
Author Response
Point 1: typo: "feed forwOrd" in Fig. 1 (p. 4).
Response 1: Thank you for your feedback. We have corrected the spelling errors in Figure 1 on page 4. Additionally, we have redrawn this figure and enhanced the clarity of the image. We apologize for these errors and appreciate your attention to detail.
Point 2: Fig. 1 that is in p. 9 should be Fig. 2.
Response 2: Thank you for your feedback. We have modified the numbering of the figure on p.9. We apologize for these errors and appreciate your attention to detail.
Thank you for taking the time to provide us with your valuable feedback. If there is anything else we can do to improve the report, please do not hesitate to let us know.