Слайд 2What’s Machine Learning?
“Machine learning will automate jobs that most people thought could
![What’s Machine Learning? “Machine learning will automate jobs that most people thought](/_ipx/f_webp&q_80&fit_contain&s_1440x1080/imagesDir/jpg/1066160/slide-1.jpg)
only be done by people.” - Dave Waters
Examples:
- Amazon recommendations based on their customer’s browsing and purchasing behavior.
- Google’s search engine, that ranks the websites by relevancy.
- Self-driving cars.
- Email spam filters.
Слайд 3Types of Machine Learning
Learning algorithms:
Supervised learning
Unsupervised learning
Reinforcement learning
Desired output:
Classification (spam, not spam)
Regression
![Types of Machine Learning Learning algorithms: Supervised learning Unsupervised learning Reinforcement learning](/_ipx/f_webp&q_80&fit_contain&s_1440x1080/imagesDir/jpg/1066160/slide-2.jpg)
(house prices, stock prices)
Clustering (customer segmentation)
Density estimation (test results of a specific number of people)
Dimensionality reduction (mapping inputs into a lower-dimensional space)
Слайд 4Neural Networks
Main objects of ANN:
Neurons (Nodes)
Layers
Learning
Connections
The goal is to mimic the way
![Neural Networks Main objects of ANN: Neurons (Nodes) Layers Learning Connections The](/_ipx/f_webp&q_80&fit_contain&s_1440x1080/imagesDir/jpg/1066160/slide-3.jpg)
the human brain organizes and make right predictions
Слайд 6Training
Back propagation:
Random weight
Take the inputs from the training set and pass them
Calculate
![Training Back propagation: Random weight Take the inputs from the training set](/_ipx/f_webp&q_80&fit_contain&s_1440x1080/imagesDir/jpg/1066160/slide-5.jpg)
error
Adjust the weights
Repeat this 10K times.
Слайд 7Training
Output = w1*in1 + w2*in2
Adustment = 0.01*err*input
![Training Output = w1*in1 + w2*in2 Adustment = 0.01*err*input](/_ipx/f_webp&q_80&fit_contain&s_1440x1080/imagesDir/jpg/1066160/slide-6.jpg)