Research map / Learning theory and algorithms / Prediction and decision losses
Adaptive Learned Bloom Filter (Ada-BF): Efficient Utilization of the Classifier with Application to Real-Time Information Filtering on the Web
Zhenwei Dai, Anshumali Shrivastava
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| divide score spectrum into more groups (larger g) in Ada-BF (ablation) | Ada-BF · Malicious URLs / Virus Scan | FPR: 25% -> 15% of LBF FPR | Improved |
| replace LBF with Ada-BF (more hash functions for high-score regions) | LBF · Malicious URLs | FPR at 500Kb: reduces FPR by 81% | Improved |
| replace LBF with disjoint Ada-BF (separate filters per group) | LBF · Malicious URLs | FPR at 500Kb: reduces FPR by 84% | Improved |
| replace LBF with Ada-BF at same FPR | LBF · Malicious URLs, FPR 0.35% | memory: 500Kb -> 300Kb | Improved |
| replace LBF with Ada-BF at same FPR | LBF · Virus Scan, FPR 0.2% | memory: 300Kb -> 150Kb | Improved |
| add sandwiched initial filter to LBF | LBF · Malicious URLs | FPR | No clear effect |
| add sandwiched initial filter to LBF at larger bitmap | LBF · Virus Scan | FPR | Improved |
| tune only c with number of groups K fixed at 10, 11 or 12, versus tuning both (K, c) (ablation) | Ada-BF and disjoint Ada-BF · Malicious URLs | FPR | No clear effect |
| use learned Bloom filters instead of standard BF | standard BF · Virus Scan | FPR: reduced by over 98% | Improved |
| use Ada-BF variants instead of BF/LBF | BF and LBF · Fake news | FPR: reduce the FPRs by over 70% | Improved |
| Ada-BF with more hash computations per query than BF | Bloom filter · Malicious URLs, 400Kb | hash operations per query: 3.39 -> 9.40 | Hurt |
Closest papers by meaning
- A Model for Learned Bloom Filters and Optimizing by Sandwiching (NeurIPS 2018 · similarity 0.89)
- Partitioned Learned Bloom Filters (ICLR 2021 · similarity 0.88)
- Ensemble Learned Bloom Filters: Two Oracles are Better than One (ICML 2025 · similarity 0.88)
- LotusFilter: Fast Diverse Nearest Neighbor Search via a Learned Cutoff Table (CVPR 2025 · similarity 0.79)
- Scaling Laws for Data Filtering-- Data Curation cannot be Compute Agnostic (CVPR 2024 · similarity 0.78)
- MExD: An Expert-Infused Diffusion Model for Whole-Slide Image Classification (CVPR 2025 · similarity 0.78)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
