Keynote: Machine Learning & Apache Spark: A Dynamic Duo
summary
Claims that IBM is the top contributor to Spark for ML and that Spark powers production use cases: a mobile app that scores building EUI from photos replaces on-site engineers using custom visual-recognition features; real-time “celebrity experience” fuses transactional and social feeds for personalized offers; Spark connectors read blockchain supply-chain ledgers to predict shipment delays; GPUs on POWER with NVLink give an order-of-magnitude performance boost for deep learning; model lifecycle includes one-click model deployment to Spark with monitoring and automated retraining.
AI-written summary · about
chapters (11)
- 0:00Introduction to ML
- 3:49Case study: building energy
- 5:19Deep learning in practice
- 6:42Personalized experiences
- 8:03Drivers of ML adoption
- 11:09The role of Apache Spark
- 12:08Contributing to open source
- 13:06Adopting spark internally
- 14:02Operationalizing models
- 15:48Blockchain and ML
- 17:02Integrating Watson
related talks
04mThe machine-learning renaissance
39mMachine Learning with Scala on Spark
17mMachine Learning at the Speed of Light with Spark
50mApache Spark vs Machine Learning
44mData Science at Scale with Apache Spark
11mHow Companies are Using Spark, and Where the Edge in Big Data Will Be
16mKeynote: Accelerating AI/ML Development with Open Source Software
22mBringing Deep Learning into the Big Data Analytics