In her talk, Zoi introduced Robopt, a novel vector-based optimizer for a cross-platform system called Rheem. Robopt uses machine learning (ML) models to predict the cost of plans in addition to scaling up the enumeration process of cross-platform query optimization via vectorization. To ease building ML models, Robopt is accompanied by a scalable training data generator. The evaluation of Robopt shows that: (i) the vector-based approach is more efficient and scalable than simply using a ML model -and- (ii) Robopt matches and, in some cases, improves Rheem’s cost-based optimizer in choosing good plans, without requiring any tuning effort. To listen to her talk, you can find a recording on YouTube.
The 36th IEEE International Conference on Data Engineering, https://www.utdallas.edu/icde/.
“ML-based Cross-Platform Query Optimization,” Zoi Kaoudi, Jorge-Arnulfo Quiané-Ruiz, Bertty Contreras-Rojas, Rodrigo Pardo-Meza, Anis Troudi, and Sanjay Chawla, https://bit.ly/2TjuG1R.
Live recordings of presentations from the ICDE Research Session No. 27 on ML and Databases 2, https://www.youtube.com/watch?v=KjIXgElcM80&feature=youtu.be.