US 12,223,408 B2Granted

Orchestrator for machine learning pipeline

AssigneeSap SE

InventorsLukas Carullo, Patrick Brose, Kun Bao, Anubhav Bhatia, Leonard Brzezinski, Lauren McMullen, Simon Lee

Life of this patent

Estimating the term…

Loading patent…

Abstract

Provided is a system and method for training and validating models in a machine learning pipeline for failure mode analytics. The machine learning pipeline may include an unsupervised training phase, a validation phase and a supervised training and scoring phase. In one example, the method may include receiving an identification of a machine learning model, executing a machine learning pipeline comprising a plurality of services which train the machine learning model via at least one of an unsupervised learning process and a supervised learning process, the machine learning pipeline being controlled by an orchestration module that triggers ordered execution of the services, and storing the trained machine learning model output from the machine learning pipeline in a database associated with the machine learning pipeline.

Claims20

Dependent claims

A computing system comprising:

a storage device configured to store a machine learning model;
and
a processor configured to:
establish a first pipeline, via a host platform, which comprises a first plurality of services for unsupervised learning of the machine learning model,
establish a second pipeline, via the host platform, which comprises a second plurality of services for supervised learning of the machine learning model, the second pipeline being independent from the first pipeline,
trigger execution of the first plurality of services via the first pipeline and train the machine learning model via unsupervised learning to generate an initially trained model,
trigger performance of one or more validation tasks for the initially trained model,
store the initially trained model in the storage device in response to performing the one or more validation tasks,
trigger execution of the second plurality of services via the second pipeline and train the initially trained model via supervised learning to generate a further trained model, and
store the further trained model in the storage device.

7 more elements

A computer-readable medium comprising instructions which when executed by a processor cause a computer to perform a method comprising:

establishing a first pipeline, via a host platform, which comprises a first plurality of services for unsupervised learning of a machine learning model;
establishing a second pipeline, via the host platform, which comprises a second plurality of services for supervised learning of the machine learning model, the second pipeline being independent from the first pipeline;
triggering execution of the first plurality of services via the first pipeline and train the machine learning model via unsupervised learning to generate an initially trained model;
triggering performance of one or more validation tasks for the initially trained model;
storing the initially trained model in the storage device in response to performing the one or more validation tasks;
triggering execution of the second plurality of services via the second pipeline and train the initially trained model via supervised learning to generate a further trained model;
and
storing the further trained model in a storage device.

5 more elements

Dependencies are inferred from the claim text.

Facts

Application: 18/111,839Filed 20 Feb 2023 · continuation
Priority: 25 Feb 2019
Published: US 2023/0206137 A129 Jun 2023
Granted: 11 Feb 2025
Term ends: Not estimatedThe source records don’t support an estimate.

Classification4 codes

G06N 20/20
Machine learning: ensemble learning
G06F 16/355
Clustering; Classification: creation or modification of classes or clusters
G06N
AI and machine learning
G06F
Computing and data processing

Family

Application16/284,291Parent of this continuation · filed 25 Feb 2019
US 2023/0206137 A1Pre-grant publication · published 29 Jun 2023
US 11,586,986 B2Parent of this continuation
US 2020/0272947 A1Published application
US 12,223,408 B2This patent

From the US filing history. Foreign family members aren’t listed.

Timeline

Citations

Open elsewhere

Building an agent or pulling many patents? Use the API or MCP server instead of scraping pages. Get an API key