A deep reinforcement learning framework with meta-learning, safety constraints, and online updating tunes learned index parameters, cutting runtime by up to 98% and raising throughput 17x in reported experiments.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.DB 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
A New Paradigm in Tuning Learned Indexes: A Reinforcement Learning Enhanced Approach
A deep reinforcement learning framework with meta-learning, safety constraints, and online updating tunes learned index parameters, cutting runtime by up to 98% and raising throughput 17x in reported experiments.