Mathematical Foundations of Poisoning Attacks on Linear Regression over Cumulative Distribution Functions
Summary: Rigorous theory of poisoning learned indexes using linear regression over CDFs: proves the known single-point attack optimal, exposes greedy multi-point attacks as suboptimal, and characterizes optimal attacks. Derives an impact upper bound that empirically nearly matches greedy loss. (summarized by gpt-5.6-luna on Jul 26 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. Atsuki Sato (University of Tokyo)
- 2. Martin Aumüller (IT University of Copenhagen)
- 3. Yusuke Matsui (University of Tokyo)
BibTeX Citation
@inproceedings{sato_sigmod26,
title = {{Mathematical Foundations of Poisoning Attacks on Linear Regression over Cumulative Distribution Functions}},
author = {Sato, Atsuki and Aumüller, Martin and Matsui, Yusuke},
series = {{SIGMOD} '26},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/3802085},
url = {https://dl.acm.org/doi/10.1145/3802085},
year = {2026}
}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 25 of 25 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 1,709 | The Tao of Inference in Privacy-Protected Databases | 2018 | VLDB |
| 2 | 12,280 | Cryptanalysis of Comparable Encryption in SIGMOD'16 | 2017 | SIGMOD |
| 3 | 11,186 | Approximating Opaque Top-k Queries | 2025 | SIGMOD |
| 4 | 10,586 | High Performance or Low Memory? An Updatable Learned Index Framework for Time-Space Tradeoff | 2026 | SIGMOD |
| 5 | 10,642 | Defense against Poisoning Attacks under Shuffle-DP | 2026 | SIGMOD |
| 6 | 10,352 | Can You Trust the Vectors in Your Vector Database? Black-Hole Attack from Embedding Space Defects | 2027 | VLDB |
| 7 | 9,677 | Robustness of Updatable Learning-based Index Advisors against Poisoning Attack | 2024 | SIGMOD |
| 8 | 8,208 | Algorithmic Complexity Attacks on Dynamic Learned Indexes | 2024 | VLDB |
| 9 | 8,808 | PACE: Poisoning Attacks on Learned Cardinality Estimation | 2024 | SIGMOD |
| 10 | 4,618 | The Price of Tailoring the Index to Your Data: Poisoning Attacks on Learned Index Structures | 2022 | SIGMOD |