Protecting Private Information for Two Classes of Aggregated Database Queries
Round 1
Reviewer 1 Report
The submitted paper belongs to one of the most important general research directions – the investigation of security and privacy. It is related to several previous articles published in Informatics and other MDPI journals recently. The present manuscript is devoted to the protection of private or confidential information from inference attacks that malicious adversaries can launch against two novel classes of aggregated database queries. These classes of queries and types of inference attacks have not been considered in the literature previously. The investigation carried out in the present paper is motivated by a fundamental result obtained by Chin and Ozsoyoglu.
The submitted paper proposes a Quadratic Audit System (QAS) protecting confidential information against Quadratic Equation Attacks (QEA), and a an Interval Audit System (IAS) protecting confidential information against an Interval Inference Attacks (IIA). Theorems 2 and 3 prove that QAS systems guarantee the protection of confidential information against QEA attacks. Theorems 4 and 5 establish that IAS systems guarantee protection against IIA attacks. In addition, Theorem 6 gives a mathematical description of matrix conditions that can be used to detect and prevent group compromise of confidential data. Proofs of all theorems are thoroughly presented. They follow all the requirements to rigorous formal proofs with high level of mathematical formalism presented and explained. Formal security proofs play central roles in modern cryptography and information science, where they are used to provide security guarantees of the systems.
It is good that the paper proposes novel systems for the protection of confidential data against new nonlinear inference attacks. These results can be used to broaden the scope of queries that can be implemented with reliable protection for the privacy. I believe that these results are excellent and can be accepted for publication in Informatics. The paper is carefully written and is easy to read. In my opinion, it can be recommended for publication in the present form.
Author Response
Response to Reviewer 1
We have made all improvements recommended by the reviewers and added thanks to three anonymous reviewers in the Acknowledgement. We added almost 4 pages of text: two sections, 19 references, 4 directions for future research, and explanations to Equations 9 and 10.
Reviewer 1, Comment 1: “I believe that these results are excellent and can be accepted for publication in Informatics. The paper is carefully written and is easy to read. In my opinion, it can be recommended for publication in the present form.”
Response: We added the following thanks to three anonymous reviewers in the Acknowledgement section: “The authors are grateful to three anonymous reviewers for thorough reports and comments that have helped to improve this paper.”
Author Response File: 
Author Response.docx
Reviewer 2 Report
Paper deals with important tasks of private information handling. The authors considered aggregate database queries.
Paper has practical value but where is the scientific novelty of this paper? It has a logical structure. The paper is well-written and technically sound. The experimental section is good.
Suggestions:
1. It would be good to add point-by-point the main contributions at the end of the Introduction section.
2. It would be good to add the remainder of this paper
3. The authors should add a strong Related works section on existing methods for solving the stated task. It would also be good to also describe non-iterative approaches for solving the stated task.
4. The paper hasn’t any Discussions.
5. Future research ideas, maybe better represented in a separate subsection of the result section.
6. The conclusion section should be extended using: the limitations of the proposed approach and prospects for future research.
Author Response
Response to Reviewer 2
We have made all improvements recommended by the reviewers and added thanks to three anonymous reviewers in the Acknowledgement. We added almost 4 pages of text: two sections, 19 references, 4 directions for future research, and explanations to Equations 9 and 10.
Reviewer 2, Comment 1: “It would be good to add point-by-point the main contributions at the end of the Introduction section.”
Response: In the end of Introduction, we added a new paragraph and a point-by-point list of the main contributions of our paper.
Comment 2: “It would be good to add the remainder of this paper.”
Response: We added new sections “Previous Work” and “Discussion”. We extended the section “Conclusion”. We added explanations to Equations 9 and 10.
Comment 3: “The authors should add a strong Related works section on existing methods for solving the stated task. It would also be good to also describe non-iterative approaches for solving the stated task.”
Response: We added new section “Previous Work”. We added several references and discussion of other alternative methods, which can also motivate suggestions for future work in the end of the paper.
Comment 4: “The paper hasn’t any Discussions.”
Response: We added new section “Discussion”.
Comment 5: “Future research ideas, maybe better represented in a separate subsection of the result section.”
Response: “”We added future research ideas in the new section “Discussion”.
Comment 6: “The conclusion section should be extended using: the limitations of the proposed approach and prospects for future research.”
Response: We have rewritten and extended the conclusion section. We added discussions of the limitations of our approach and 4 directions for future research into the new section Discussion.
Author Response File: Author Response.docx
Reviewer 3 Report
In this paper, the authors proposed a novel private information protection method for two classes of aggregated database queries. The paper is logically clear and easy to follow.
However, there are some issues to be fixed.
1) The contribution of this work is not fully highlighted, which makes it a bit unclear.
2) The authors are expected to explore existing literature to distinguish the proposed idea with other relevant research. A related work section would be helpful in this case. Consider discussing differential privacy (Customizable Reliable Privacy-Preserving Data Sharing in Cyber-Physical Social Network) and federated learning (FedTwin: Blockchain-Enabled Adaptive Asynchronous Federated Learning for Digital Twin Networks) as well since both techniques can serve this topic as well.
3) The authors may consider better explain the equations, for example, Equation 10.
4) Some evaluation results would be really helpful to justify the proposed idea.
5) It would be better if using a table to illustrate the abbreviations on Page 16.
6) Typos, wording issues, and grammar errors should be further taken care of.
Author Response
Response to Reviewer 3
We have made all improvements recommended by the reviewers and added thanks to three anonymous reviewers in the Acknowledgement. We added almost 4 pages of text: two sections, 19 references, 4 directions for future research, and explanations to Equations 9 and 10.
Reviewer 3, Comment 1: “The contribution of this work is not fully highlighted, which makes it a bit unclear.”
Response: In the end of Introduction, we added a new paragraph and a point-by-point list of the main contributions of our paper.
Comment 2: “The authors are expected to explore existing literature to distinguish the proposed idea with other relevant research. A related work section would be helpful in this case. Consider discussing differential privacy (Customizable Reliable Privacy-Preserving Data Sharing in Cyber-Physical Social Network) and federated learning (FedTwin: Blockchain-Enabled Adaptive Asynchronous Federated Learning for Digital Twin Networks) as well since both techniques can serve this topic as well.”
Response: We added new section “Previous Work”. We added several references and brief discussions of differential privacy and federated learning as recommended.
Comment 3: “The authors may consider better explain the equations, for example, Equation 10.”
Response: We have rewritten the whole paragraph explaining Equation 10 and added new sentences. We added notation E() for expected value to Equations 9 and 10, since it makes the formulas clearer. We added two references where the readers can find a complete proof of Equation 10, additional explanations and worked examples.
Comment 4: “Some evaluation results would be really helpful to justify the proposed idea.”
Response: We added this suggestion as Direction 3 for future research in the new discussion section. It would be best to consider this in separate future papers in conjunction with Direction 2.
Comment 5: “It would be better if using a table to illustrate the abbreviations on Page 16.”
Response: We added Table 2 to present the abbreviations and subsections, where they are explained. Our file uses the MDPI LaTeX command \abbreviations{Abbreviations}{} as indicated in the MDPI template.
Comment 6: “Typos, wording issues, and grammar errors should be further taken care of.”
Response: After the major revisions, we carefully reviewed the whole text. Then we proofread the text again to make sure that there are no spelling or grammar errors.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The authors took into account all my comments and accordingly improved the text of the article, which is now of better quality and significantly improved. My recommendation is to accept this paper for publishing as it is.
Reviewer 3 Report
After reading through the revised manuscript, the reviewer believes the authors have addressed all my concerns. The paper is in good in shape and logically clear. The novelty, significance, soundness, etc., are at a satisfying level. The reviewer recommends the paper to be accepted for publication.
