Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Electrical and Electronic Engineering | en_US |
| dc.contributor.advisor | Ye, Qingqing (EEE) | en_US |
| dc.creator | Ding, Meng | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/13907 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | en_US |
| dc.rights | All rights reserved | en_US |
| dc.title | Accurate graph statistics estimation with differential privacy | en_US |
| dcterms.abstract | Triangle counting is one of the fundamental tasks in subgraph analysis, as it helps reveal the connectivity among three nodes (users). In recent years, increasing efforts have been made to apply the LDP (Local Differential Privacy) model to this task, in order to provide strong privacy guarantees – even against the data collector (i.e., the curator). In this paper, we consider the structure 2-star, which containing richer information than adjacency bit by encoding the connectivity of two edges simultaneously, and aim to improve the accuracy of triangle counting under LDP. | en_US |
| dcterms.abstract | In our initial attempt, each user perturbs and reports its complete 2-star vector, where each 2-star is centered at itself. However, we find that this approach is ineffective, due to the high sensitivity of the local function which outputs the complete 2-star vector. To address this issue, we adopt the sampling approach introduced by Imola et al., which dramatically reduces sensitivity. However, their triangle counting algorithm suffers from relatively large sampling errors. Therefore, in our second scheme, we introduce a relaxed user pairing approach, combined with a selective clipping strategy, to help balance the sampling error and noise error. Under the same setting of shuffle model (close to LDP with additional assumptions), by experiments we show that even in a specific case, our counting algorithm improves accuracy across a range of commonly used total privacy budget values. | en_US |
| dcterms.extent | vi, 48 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2025 | en_US |
| dcterms.educationalLevel | M.Sc. | en_US |
| dcterms.educationalLevel | All Master | en_US |
| dcterms.accessRights | restricted access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 8316.pdf | For All Users (off-campus access for PolyU Staff & Students only) | 689.6 kB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/13907