A classifier that skips Bloom filter probes can halve measured GET latency in an LSM-tree but introduces false negatives, while a learned Bloom filter cuts per-level memory 70-80% with zero false negatives in static tests.
A model for learned bloom filters and optimizing by sandwiching
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DS 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Learned LSM-trees: Two Approaches Using Learned Bloom Filters
A classifier that skips Bloom filter probes can halve measured GET latency in an LSM-tree but introduces false negatives, while a learned Bloom filter cuts per-level memory 70-80% with zero false negatives in static tests.