Position Distribution Matters: A Graph-Based Binary Function Similarity Analysis Method

Round 1

Reviewer 1 Report

This study proposes a graph-based method in a 3-step style for binary function similarity analysis. The topic is interesting and noteworthy. The manuscript is well written so only minor modifications are required. The contribution presented in the introduction needs to be extended and discussed in the conclusion too. In the case of Chapter 5, it is recommended to move as it tends to obstruct the flow of the manuscript. Lastly, check the typos again for the reader.

Author Response

Thanks very much for your comments. We have added part of the contribution in discussion chapter and highlight the changes.The original Chapter 5 (Related Works) has been moved to Chapter 2 for adjusting the whole structure. We have also checked the typos in this paper.

Reviewer 2 Report

- reference 18, 34 should be replaced/removed because there are old/outdated. Also, reference 2 has to be to be completed with more information in References chapter (e.g. source, date).

- the Related work chapter has to be moved before chapter 3, maybe before chapter 2.

- the English must be improved in the paper.

- generally, all chapters are justified and the flow of ideas is clear. The purpose, motivation, methodology, case-study and results, are well-presented. The paper is not exceeding in complexity, but it is a good work.

Author Response

Thanks very much for your comments, and the responses are as below:

• We have deleted old references 18 and 34, and added information for reference 2.

• The Related Work chapter has been moved to be Chapter 2 for better structure.

• We have checked the typos and improved the English writing.

Reviewer 3 Report

The authors propose in this paper a graph-based binary function similarity detection method. The methodology, following the 3-step approach, as the authors indicate, is rigorously explained. On the other hand, the experiments carried out are successful and, according to the results observed, the proposed method improves other techniques in similarity detection and vulnerability detection tasks.

The authors are also aware of the limitations of their method, especially when employing these learning-based techniques.

Regarding the structure of the article, it is correct, as well as the style used.

However, it is noted that figures 1 and 3 should improve the quality, especially the size of the text embedded in them. Finally, authors are advised to review the style and format of references 11-18. It would also be interesting to know if there is more recent literature than the one that appears in the references since the dates of the works cited are a bit old.

Author Response

Thanks very much for your comments.

We have fixed figures 1 and 3 by adding correct sub captions, instead of embedding text in the figure.

We also removed some old references, which was added mistakenly to supplement the content, and other related works we cited are carefully checked to be recent and latest.