5 Levels of AI in Microsoft Azure
summary
Practical walkthrough of Azure tools with measured results: Azure Video Indexer produces transcripts, translations and celebrity tags; Custom Vision trains object detectors with ~100 images per class and hit mean average precision ~85% for fruit and correctly labeled a Tesla Model S at 96% confidence; AutoML ran an apartment price experiment (one run ~1h03) and deployed a real-time endpoint predicting ~4.5M SEK for a sample input; and a Annoy image search built from ResNet18 feature vectors runs as an Azure ML job in ~5 minutes costing under $0.01.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker background
- 5:16Five levels of AI explained
- 8:27AI applications with Video Indexer
- 12:41Cognitive Services demo and overview
- 15:48Custom Vision and model training
- 24:23Azure ML for team-based development
- 38:46Machine learning frameworks
- 51:08Advanced AI with algorithms
- 59:20Q&A and closing remarks
related talks
57mEasier AI allows everyone to build smart apps
46mBuilding accessible experiences with AI
50mAzure AI, Power new possibilities for every organization
01h 03mEasier AI allows everyone to build smart apps
01h 00mLearn Live: Get started with AI on Azure
16mHow to get started with Computer Vision AutoML in Azure
44mThe age of applied AI
32mUnleash your AI potential