Three Key Types of Machine Learning

• Supervised Machine Learning

• Unsupervised Machine Learning

• Automated Machine Learning

Supervised Machine Learning

• User provides labeled data

• Computer analyzes the provided data to predict labels

• Hands on

Unsupervised Machine Learning

• Raw, unlabeled data is provided

• Algorithms detect patterns

• Hands off

Automated Machine Learning

• Raw data is entered into a software program

• Algorithm adapts to the data

• User specifies parameters for repetition

• Program develops optimum models to fit the data

Use Cases for Automated Machine Learning

Apple and Google Photos

• User adds photos to camera roll

• Algorithms identify objects, animals, people, and landscape features

• Program identifies people from images that have not been explored yet.

Supervision

• User identifies specific people by name

• When people are detected, they are labeled with the provided name

Use Cases for Automated Machine Learning

Google Search

• Search Bar text prediction

– Each search is a new piece of data

– Software analyzes searches in real time

– Recommended searches are provided based on previous searches

Resources for Further Exploration

• Microsoft Azure

https://docs.microsoft.com/en-us/azure/machine-learning/concept-automated-ml

• IBM Cloud

https://www.ibm.com/cloud/learn/machine-learning#toc-how-machin-NoVMSZI_

• Automated Machine Learning Deep Dive

https://library.oapen.org/viewer/web/viewer.html?file=/bitstream/handle/20.500.12657/23012/1007149.pdf?sequence=1&isAllowed=y

Works Cited

• “What Is Automated ML? Automl - Azure Machine Learning.” What Is Automated ML? AutoML - Azure Machine Learning | Microsoft Docs, https://docs.microsoft.com/en-us/azure/machine-learning/concept-automated-ml.

• Hutter, Frank, Kotthoff, Lars and Vanschoren, Joaquin. Automated Machine Learning. Springer, 2019.