BoomHQ learns to rewrite hybrid vector-scalar queries using autoencoder-modeled correlations and neighborhood selectivity estimates, delivering 2x average and 25x peak speedups at target recall levels across three vector DBMSs.
Advances in neural information processing Systems32(2019)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
BoomHQ: Learning to Boost Multiple Hybrid Queries on Vector DBMSs
BoomHQ learns to rewrite hybrid vector-scalar queries using autoencoder-modeled correlations and neighborhood selectivity estimates, delivering 2x average and 25x peak speedups at target recall levels across three vector DBMSs.