Predictive Maintenance
summary
Claims production ML models predict maintenance within three months for ASML lithography machines using component-level classifiers and transfer and active learning to adapt across products. Data comes from 15,000 sensors streaming 4.5 terabits per day. Deployment uses Google Cloud (DataProc, container registry, Cloud Build) with CI/CD + continuous training pipelines, a label‑inspection loop and feature‑importance outputs. Result: tens of models in production and a reported reduction in time to market about eightfold.
AI-written summary · about
chapters (5)
- 0:00Start
- 1:01Introduction to ASML
- 14:00Model Validation and Explainability
- 15:31Predictive Maintenance Model Creation Cycle
- 18:40Machine Learning Operations (MLOps)
related talks
14mTransforming Manufacturing: AI-Driven Predictive Maintenance
43mPredictive Maintenance: From POC to Production With Spark
26mProactive Failure Prediction Through Anomaly Detection
42mImproving Machine Learning Development Reliability
26mPredicting the future (of equipment) using ML.NET
19mEnterprise AI-Powered Predictive Maintenance Platform
22mEngineering Predictive Platforms
01h 05mMachine Learning Models and Productionization