Understanding Robustness Issues of Updatable Learned Indexes: [Experiments & Analysis]
Summary: Systematic benchmarking shows updatable learned indexes lack robustness: real-time model instability erodes gains and they rarely beat traditional indexes in workloads. Root causes include overfitting and unbalanced structures; mitigations offered. (summarized by gpt-5-nano on Feb 09 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Yuanhui Luo (Renmin University of China)
- 2. Minhui Xie (Renmin University of China)
- 3. Yiheng Tong (Renmin University of China)
- 4. Shichao Jiang (Renmin University of China)
- 5. Yunpeng Chai (Renmin University of China)
BibTeX Citation
@inproceedings{luo_sigmod26,
title = {{Understanding Robustness Issues of Updatable Learned Indexes: [Experiments \& Analysis]}},
author = {Luo, Yuanhui and Xie, Minhui and Tong, Yiheng and Jiang, Shichao and Chai, Yunpeng},
series = {{SIGMOD} '26},
booktitle = {Proceedings of the {ACM} {SIGMOD} International Conference on Management of Data},
publisher = {Association for Computing Machinery},
doi = {10.1145/3749188},
url = {https://dl.acm.org/doi/10.1145/3749188},
year = {2026}
}
Incoming Citations (Sorted by Pagerank)
Showing 2 of 2 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 10,272 | NeurBench: A Benchmark Suite for Learned Database Components with Drift Modeling: [Experiments & Analysis] | 2026 | SIGMOD | 5.093636e-05 |
| 10,277 | On Self-Designing Learned Indexes | 2026 | SIGMOD | 5.093636e-05 |
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 34 of 34 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 | 10,277 | On Self-Designing Learned Indexes | 2026 | SIGMOD |
| 2 | 9,490 | Robustness of Updatable Learning-based Index Advisors against Poisoning Attack | 2024 | SIGMOD |
| 3 | 6,687 | Making In-Memory Learned Indexes Efficient on Disk | 2024 | SIGMOD |
| 4 | 43 | The Case for Learned Index Structures | 2018 | SIGMOD |
| 5 | 1,551 | Updatable Learned Index with Precise Positions | 2021 | VLDB |
| 6 | 10,378 | High Performance or Low Memory? An Updatable Learned Index Framework for Time-Space Tradeoff | 2026 | SIGMOD |
| 7 | 5,423 | Updatable Learned Indexes Meet Disk-Resident DBMS - From Evaluations to Design Choices | 2023 | SIGMOD |
| 8 | 3,792 | Learned Index: A Comprehensive Experimental Evaluation | 2023 | VLDB |
| 9 | 847 | Benchmarking Learned Indexes | 2021 | VLDB |
| 10 | 2,806 | Are Updatable Learned Indexes Ready? | 2022 | VLDB |