Ckdew

joined 1 year ago
[–] Ckdew@alien.top 1 points 1 year ago

I guess my point is, causal inference has been around for a long time whereas Causal ML just popped up a few years ago. So how is Causal ML different from Causal inference. What are the unique problem Causal ML tries to solve that Causal inference couldn't?

 

I'm currently doing an Master in Economics with a focus on Econometrics. I've taken Calc 1-3, Linear Algebra, Real Analysis. Though I haven't taken a formal Probability theory/mathematical statistics class, I do use them implicitly in my graduate level Econometrics course.

I would greatly appreciate advice on which of the following courses I should take to transition into ML research:

  1. Measure theory: Probability spaces and probability measures. Random variables. Expectation and integration. Convergence of random variables. Conditional expectation. The Radon-Nikodym Theorem. Martingales. Stochastic processes. Brownian motion. The Itô integral
  2. Stochastic Processes: The course examines Martingales, Poisson Processes, Brownian motion, stochastic differential equations and diffusion processes.
  3. Time series: autocorrelation; stationarity; causality;  basic time series models: AR, MA, ARMA; ARCH and GARCH models for financial time series; trend removal and seasonal adjustment; invertibility; spectral analysis; estimation; forecasting. We will also discuss nonstationarity and multivariate time series.

Ps I can also take ML courses but I thought these courses would allow me to build a stronger foundation.

Not sure which areas of ML I would specialise in but it'll likely be Causal ML given my Economics background

 

Looks like classic causal inference, the type would see in Pearl or Econometrics. I don't understand where the ML comes in.