Taking Models to the Next Level with Azure Machine Learning
summary
Tailwind Traders moves from Azure Cognitive Services Text Analytics to a bespoke pipeline using Azure Machine Learning and the Azure ML SDK, running Automated ML on ~10,000 product reviews (autoML reports 0.97 accuracy) and then training an open‑source nlp architect aspect‑based sentiment model on CPU/GPU compute clusters. The run produces aspect and opinion lexicons, logs stream back to the notebook, and the trained model and scoring script are saved for deployment; hyperparameter search and distributed compute are used and the workflow works end‑to‑end.
AI-written summary · about
chapters (30)
- 0:00Start
- 0:45Resources
- 2:13Some Basic Machine Learning Applications
- 3:17Language Understanding
- 4:10Unlocking Unstructured Data
- 6:11Cognitive Services
- 8:53The 80 20 Rule
- 9:44Downsides
- 10:14Managing Machine Learning Models
- 11:32Azure Machine Learning
- 13:15What Is Automated Machine Learning
- 15:23Azure Machine Learning Studio
- 15:34Azure Machine Learning Studio
- 15:54Create a Baseline Model
- 16:07Data Sets
- 18:08Configure the Run
- 19:27Select the Type of Machine Learning Task
- 20:40Exit Criterion
- 20:59Metric Threshold
- 21:23Validation
- 24:53Automated Machine Learning
- 28:09Environmental Dependencies
- 28:53Data and Compute Management
- 29:46Distributed Training
- 35:09Data Science Tools
- 37:15Set Up Our Environment
- 37:41Attach Our Experiment to Our Compute Cluster
- 39:19Creating a an Estimator
- 44:44Submit this Experiment
- 49:39The Azure Machine Learning Learning Paths
related talks
57mTaking Models to the Next Level with Azure Machine Learning
58mDeep Dive into Azure Automated Machine Learning
01h 14mAI of Things
01h 05mAzure Machine Learning for the Developer
41mLeveraging Automated Machine Learning to enable anyone to develop Machine Learning solutions
21mFast and Efficent Text Analysis using Azure's Language Cognitive Service
01h 01mAI and ML imbedded in Azure Data Platform
49mAutomated Machine Learning on Azure