Edward Screven on machine learning
From Oracle Live: MySQL HeatWave—New Machine Learning Capabilities | FULL SHOW · · Oracle
“Machine learning is built into MySQL HeatWave — model building, hyperparameter selection, training and inference all happen inside the database; the marginal cost of HeatWave ML is zero because it's included.”
On , Edward Screven, Executive Vice President & Chief Corporate Architect at Oracle Corporation, spoke about machine learning during Oracle Live: MySQL HeatWave—New Machine Learning Capabilities | FULL SHOW on Oracle.
About Edward Screven
Edward Screven, Oracle's chief corporate architect, has been promoting the company's MySQL HeatWave service, which he described as one of the fastest-growing OCI services since its introduction in October 2020. In a January 2023 presentation, Screven announced new machine learning capabilities for MySQL HeatWave, stating that model building, hyperparameter selection, training, and inference all occur within the database. He said the service allows users to perform machine learning without needing to be a data scientist and that the marginal cost of HeatWave ML is zero because it is included. Screven contrasted the service with competitors, saying that Redshift's AutoML exposes hyperparameters that shift the tuning burden to the user, while MySQL HeatWave's AutoML requires no such configuration. Screven has also discussed Oracle's internal use of its own cloud infrastructure. In a December 2021 presentation, he said Oracle runs its entire business on its own cloud, generating $41 billion in revenue annually. He stated that the company's migration to Oracle Cloud Infrastructure reduced compute requirements by 50%, increased performance by 30%, and reduced system provisioning from weeks to minutes. Additionally, Screven has spoken about Oracle Enterprise Manager 13c Release 5, which he said is designed for customers with both on-premise and cloud infrastructure, allowing them to manage both environments with a single tool. He cited a model from Peak Solutions that estimated a 60% reduction in operating costs for a hypothetical customer using Enterprise Manager.