Here's a great ML interview question:
First ask about regularization. The candidate will likely talk about norm penalties on weights.
Then ask "why do we do lambda norm penalties, - lambda|w| , instead of directly constraining |w|<R if they're essentially equivalent under KKT?"
There are three really nice perspectives
1. Optimization theory
2. Bayesian theory
3. ER learning theory
Stock Prediction AI: Using Machine Learning and Deep Learning to predict stock price movements in Python.
The Python code is 100% free on GitHub.
Let's dive in (bookmark this):
Which programming channel do you follow for which content?
As someone who's watched many, I'd like to see what you are seeing, and I'd tell you if it sucks, and provide better options (IMO).