Machine Learning explained with Examples
summary
Raspberry Pi face‑recognition prototype actually works out of the box but misclassifies a book cover and carpet as faces; a quick hack required no training. An Oracle in‑database drag‑and‑drop ML workflow shows classification with decision trees, plus autonomous data warehouse + Apache Zeppelin notebooks for exploration. A beer‑rating demo predicts probability of a 5, telco churn experiments run in Python, and the claim that 80% of ML is data prep recurs; Facebook runs production with a feature store, FB Learner Flow and a predictor, using PyTorch/Caffe2 tooling.
AI-written summary · about
chapters (22)
- 0:00Intro
- 0:15Introduction, Heli
- 1:50What is Machine Learning?
- 4:22A simple example, Chatbot
- 6:16Real life use cases for ML
- 10:28Second Test
- 11:30In database Machine Learning
- 11:47Why in database?
- 14:26Oracle SQL Developer, Data Miner
- 15:45Oracle SOL Developer demo
- 19:02SQL Oracle Autonomous DW
- 33:20Facebook's mission
- 34:56News Feed
- 36:28Predicting the Clicks
- 40:41Language Translation
- 41:35Speech Recognition
- 43:42FBLearner Platform
- 45:36Frameworks for deep learning
- 46:58Open Neural Network Exchange, ONNX
- 48:04The success factors 12
- 49:28What to measure?
- 52:21Conclusion
related talks
51mMachine Learning Exposed Introduction to Machine Learning
50mBuilding Intelligent Mobile Apps using Machine Learning
22mReal-world applications of machine learning: Beyond the hype
55mMachine Learning and AI in Java on AWS
41mBeyond the Algorithm: What Makes Machine Learning Work?
42mMachine Learning Crash Course
52mMachine Learning Crash Course
35mIntelligent Ruby: Getting Started with Machine Learning