Deep Learning in Microsoft Azure: CNTK, CaffeOnSpark and Tensorflow
summary
Microsoft positions deep learning as an ecosystem: deploy models in Azure Notebooks and Azure Machine Learning workspaces, train with TensorFlow, CNTK or CaffeOnSpark on HDInsight, export as ONNX and serve via ML.NET or cognitive services. Demo shows AutoML selecting a classifier (MNIST sample, ~0.99 accuracy) within seconds and Azure-hosted pipelines that push models to edge or cloud. Real-world claim: ABB reports a 20% boost in customer satisfaction by using Azure cognitive services.
AI-written summary · about
chapters (9)
- 0:00Introduction to AI
- 2:50Defining AI and machine learning
- 5:56Use cases for AI
- 8:20Data-driven projects
- 12:39Starting with AI
- 17:36Azure notebooks overview
- 26:04Demo: Training with AutoML
- 35:34ML.NET and tools
- 49:42Deep learning and conclusion
related talks
01h 01mAI and ML imbedded in Azure Data Platform
34mMaking Data Scientists Productive in Azure
31mAzure AI Innovations at Microsoft
33mBuilding end-to-end computer vision solutions from deep learning models
24mDon't want to miss out on the AI wave?
33mApplied AI: Real-world use cases for Microsoft's Azure Cognitive Services
58mDeep Dive into Azure Automated Machine Learning
24mHow to build next-gen AI services with NVIDIA AI on Azure Cloud