Machine Learning Introduction
Hack University Machine Learning Introduction
- 25 min talk
- 5 min break
- 25 min project
5 minute break
- What is Machine Learning
- Machine Learning Applications
- Linear Regression
- Temperature Conversion
- Artificial Intelligence?
- Modeling (Regression)?
- Data Science?
You signed up for this, what do you think?
- Machine learning is like automated Data Science
- But you can’t automate it all
Machine’s need a lot of help
- In future sessions we’ll learn about generative models
- For now, Data Scientists still have jobs
It takes human understanding to build a system
(Otherwise you wouldn’t be here)
And it requires a lot of effort to get data ready for machine learning
So maybe Machine Learning is:
What a software developer would do if she were given a data science problem.
Machine Learning expert:
Someone that knows more statistics than a developer and more computer science than a statistician”
A Data Scientist is
“A statistician in Silicon Valley”
- You’re going to learn statistics if you are a python developer
- You’re going to learn python and data if you already know stats
- You’re going to to know a lot about everything when we’re done
What do you want to learn?
- Tell me about your plans, career, ideas
Using data to create representations or abstractions of data.
- Sometimes representations are useful for predicting
This is what most experts think of when they think of Machine Learning…
- Machines interacting with the world based on predictions
- Agents trying to
- achieve a goal
- optimize something
- playing Breakout
- playing Galaga
- playing Go
Predicting the future is what humans spend their lives learning how to do. It’s why children perform “experiments” by pushing their food of the table and onto the floor. They are testing their “hypothesis” that a caregiver will put it back on the table or clean up the mess. They are also testing their hypothesis that the food won’t just disappear, or transform into something different (like icecream), or become unedible.