Integrating query optimization with machine learning model prediction
AssigneeAmazon Technologies, Inc.
InventorsBalakrishnan Narayanaswamy, Gokul Soundararajan, Jiayuan Chen, Yannis Papakonstantinou, Vuk Ercegovac, George Constantin Caragea, Sriram Krishnamurthy, Nikolaos Koulouris
Life of this patent
Estimating the term…
Abstract
A database system may include a machine learning model which may be used to perform various data analytics for data stored in the database system. In response to a request to invoke the machine learning model to generate a prediction from data stored in the database system, the database system may perform one or more optimization operations, as part of a query plan, to prepare the data to make it suitable for use by the machine learning model.
Claims20
A system, comprising:
- one or more hardware processors and memory configured to implement one or more computing resources of a database system including a query engine that includes one or more of the computing resources of the database system,
- wherein the query engine is configured to:
- receive a statement in a query language supported by the database system that specifies both a predicate to obtain a first set of data stored in the database system and a machine learning model, which causes the database system to determine one or more operations from the statement to generate a prediction for the first set of data stored in the database system using the machine learning model;
- obtain the first set of data from the database system;
- perform an operation to prepare the first set of data for use by the machine learning model to generate the prediction;
- and
- apply the machine learning model to the prepared first set of data at the one or more computing resources of the query engine to generate the prediction for the first set of data stored in the database system.
A method, comprising:
- receiving, at a query engine including one or more computing resources of a database system, a statement in a query language supported by the database system that specifies both a predicate to obtain a first set of data stored in the database system and a machine learning model, which causes the database system to determine one or more operations from the statement to generate a prediction for the first set of data stored in the database system using the machine learning model;
- performing, at the one or more computing resources of the query engine, an operation to prepare the first set of data for use by the machine learning model to generate the prediction;
- and
- applying, at the one or more computing resources of the query engine, the machine learning model to the prepared first set of data to generate the prediction for the first set of data stored in the database system.
One or more non-transitory computer readable media storing instructions that, when executed on or across one or more processors, cause the one or more processors to implement:
- responsive to receiving a statement in a query language supported by a database system via a network interface, wherein the statement specifies both a predicate to obtain a first set of data stored in the database system and a machine learning model, which causes the database system to determine one or more operations from the statement to generate a prediction for the first set of data stored in the database system using the machine learning model,
- performing, at one or more computing resources selected out of a plurality of computing resources by the database system to process statements to the database system, an operation to prepare the first set of data for use by the machine learning model to generate the prediction;
- and
- applying the machine learning model to the prepared first set of data at the selected one or more computing resources to generate the prediction for the first set of data stored in the database system.
Dependencies are inferred from the claim text.