Readings on causality

Post date: Jul 21, 2016 9:53:38 PM

July 20

Decision led by David Lopez Paz

- can we mix experimental and observational data?

- can we see that training and test follow the same distribution?

- what guarantees of learnability and how does the learnability affect the validity of the results?

- the method of David et al can be thought of as density estimation in the first layer

- we should talk more about the generating process of the CE pairs