This week I want you to give us some understanding of how machine learning works. Here are some of the kinds of questions that you might be able to answer:
- What are the names of the algorithms?
- Which algorithm will help us solve which kind of problem?
- Are there cautions that we should exercise when using a particular algorithm?
- Should we avoid the use of some particular algorithms in some circumstances?
- What are the big ideas in these algorithms?
- Are some algorithms related to one another?
- Do they share common features?
- What kind of software will I want to use in my explorations of machine learning?
Look hard and long on the Web for tutorials. Try many searches on the Web and in the library's databases. Try searching with many different keywords.
Budget many hours for your search. The first items that come up are unlikely to meet your needs this time.
Allow for time to comprehend what you find. A single reading, viewing, or listening is unlikely to be sufficient.
Be brave. Try reading articles and listening to videos and podcasts that contain challenging content. Much of it will go over your head. Still, you can learn something. No matter how far you go in your studies, you will always find difficult articles, books, and lectures. Can you find an illustration, a statistic, or a paragraph in the introduction or conclusion of a technical article that you can understand?
Be creative. We began two weeks ago by listing keywords that we might use in searching for articles about machine learning. In your reading, you have certainly encountered more words and names. Search for authors who have written overviews and introductions for people with your skills and knowledge.
Here are some of the words we heard in today's video:
- linear regression
- gradient descent
- naive Bayes
- decision tree
- logistic regression
- log-loss function
- neural network
- linear optimization
- support vector machine
- kernel trick
- k-means clustering
- hierarchical clustering
Do searches with any of these words yield useful results? Do searches with the names of authors, teachers, presenters, and organizations that you see listed in one resource help you find other resources?
You might find some of what you need by returning to resources that you and your classmates previously identified.
Share what you find with your classmates. Ask classmates what they are finding. Help one another.