Apr 8, 2020 · Bao combines modern tree convolutional neural networks with Thompson sampling, a decades-old and well-studied reinforcement learning algorithm.

Jun 1, 2022 · We demonstrate that Bao can quickly learn strategies that improve end-to-end query execution performance, including tail latency, for several workloads ...

Thompson Sampling. Balances between exploration and exploitation when training. Exploration only: choose actions at random.

Bao is a learned query optimizer that learns to "steer" the PostgreSQL optimizer by issuing coarse-grained query hints. For more information about Bao ...

• Bao: “first learned query optimisation system that outperforms both open source and ... Bao:Learning to Steer Query Optimizers.2020. arXiv preprint https ...

May 10, 2022 · This work introduced Bao, a bandit optimizer which steers a query optimizer using reinforcement learning. Bao is ca- pable of matching the ...

Jun 25, 2021 · This work introduced Bao, a bandit optimizer which steers a query optimizer using reinforcement learning. Bao is capable of matching the ...

Sep 10, 2024 · Bao combines modern tree convolutional neural networks with Thompson sampling, a decades-old and well-studied reinforcement learning algorithm.

Bao takes advantage of the wisdom built into existing query optimizers by providing per-query optimization hints, and combines modern tree convolutional ...

Bao: Introduction. Bao: Bandit optimizer. By steering a traditional query optimizer, Bao: → Outperforms PG after 1 hour of training. → Reduces 99% latency.