Week 12 - AutoML

Question 1

What is autoML

Answer 1

Automated Machine Learning (AutoML) is the process of automating the end-to-end process of applying machine learning to real-world problems, including data preprocessing, feature selection, model selection, and hyperparameter tuning, to make ML accessible and efficient.

Question 2

What are the benefits of grid search

Answer 2

+ very simple approach +
+ can be used to study the problem +
- does not scale to high dimentions -
- grid needs to be defined -

Question 3

What are the advantages and disdvantage of random search

Answer 3

Not data efficient
computationally expensive

Even more simple
easily parallelizable

Question 4

what are the advantages and disadvantages of bayesian optimization

Answer 4

not easily to parallelize

Data efficient

Question 5

What are the 4 steps in Beyesian Optimisation

Answer 5

1. Fit a probabilistic model on observations
2. Compute acquisition function to find promising points
3. Evaluate next point
4. Repeat

Question 6

give 3 acquisition functions

Answer 6

• expected improvement
• upper condidence bound
• Entropy search