Paths to anomaly detection w/ TIBCO data science, streaming on AWS
summary
Demo shows TIBCO Data Science and TIBCO Streaming running on AWS (EMR, SageMaker, Kinesis, Spotfire) that detects a Ring of Fire wafer defect by vectorizing wafer maps (Bessel radial bases), using SVD + k‑means and deep autoencoder reconstructions, and deploying the classifier to the event stream for real‑time labeling. Pipeline reduces ~1,000,000 process variables to ~50 candidates via SAX encoding and wide‑data correlation; a refresh of the digital‑twin regression runs in ~15–50 minutes and surfaces root causes for rapid action.
AI-written summary · about
chapters (17)
- 0:00Intro
- 0:28The Ideal Data Science Platform
- 5:14Data Science in the Cloud: Leidas Healthcare Analytics
- 7:02TIBCO analytics transformation platform
- 10:28TIBCO Data Science and AWS
- 12:49TIBCO Data Science Solutions on AWS
- 14:18Anomaly Analysis Solution Overview
- 14:43Data Challenges in High-Tech Manufacturing
- 15:52Cross-Sectional Anomaly Analysis
- 16:10Spatial Anomaly Analysis
- 17:10Cross-Sectional Anomaly Detection
- 28:50Anomaly Detection and Analysis
- 30:54The Extreme Challenge of Big & Wide Data
- 39:02Solution Architecture
- 40:51Performance Benchmarks & Conclusions
- 42:33Mueen's Algorithm for Similarity Search
- 45:00Manufacturing Batches
related talks
29mAI-Powered Anomaly Detection in Cloud Data Pipelines
44mAnomaly Detection in Telecom with Spark
39mMastering real-time anomaly detection with open source tools
29mAnomaly Detection
01h 01mBetter reliability and uptime with analytics and observability
19mBig Data
31mAccelerating Your Analytics Journey With Real-time AI
30mAnomaly Detection with Apache Pinot & ThirdEye