From Zero to Hero: Detecting Leaked Data through Synthetic Data Injection and Model Querying
Summary: Injects a small set of synthetic records with localized class‑distribution shifts into an owner's tabular dataset so models trained on leaked data exhibit distinctive prediction patterns under query. LDSS is model‑oblivious, works for classification (and regression), and is robust across models/datasets. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Biao Wu
- 2. Qiang Huang
- 3. Anthony K. H. Tung
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 2 of 2 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 4,243 | Locality-Sensitive Hashing Scheme based on Longest Circular Co-Substring | 2020 | SIGMOD | 6.32976e-05 |
| 5,456 | Point-to-Hyperplane Nearest Neighbor Search Beyond the Unit Hypersphere | 2021 | SIGMOD | 5.4976692e-05 |
Previous
Page 1 / 1
Next