Can you solve it?
You get two datasets, but...
...they are fully de-identified.
Your client wants to know the effect of one of their marketing activities (X) on a sales-related outcome (Y).
🧵 1/
#CausalSky #EpiSky #EconSky
Life update: I successfully defended my Ph.D. last week and soon I will be able to introduce myself like James Bond
(I've always wanted to do this... 😅)
giphy.com/gifs/fFEdJfI...
As seem on the internet:
"Academia is wild - the socially isolated recluse who avoids eye contact has to teach a classroom of 22 year olds"
How to improve the Pixar logo.
📹 theoretically.teddy
Meanwhile I am without an official address and resultantly many of my administrative processes are either on hold or terminated.
A German colleague at work summed it up nicely today: "I wonder why Germany keeps saying we need to attract top talent and then treats them like garbage once they arrive"
#Rant
3 months after relocating, I still wait for appointment to register my new address.
The People's office has a plethora of such appointments, but I must wait for the "Foreigners' office" to call me.
All this despite having a permanent residence and 7+ years of paying same taxes in Germany. 🥲
Small sample experiments are useless Unless... In his new blog post, Mårten Schultzberg explains how Spotify leverages "smaller sample" experiments to improve product-related decision-making Mårten defines "small sample" as a sample that "doesn't let you estimate.. 1/ 👇🏼 #CausalSky #ABtesting
Small sample experiments are useless
Unless...
In his new blog post, Mårten Schultzberg explains how Spotify leverages "smaller sample" experiments to improve product-related decision-making
Mårten defines "small sample" as a sample that "doesn't let you estimate..
1/ 👇🏼
#CausalSky #ABtesting
The belief that causal methods are intellectually interesting but practically too difficult stops many people from using them.
Eyal Kazin, PhD @zur-do.bsky.social (Zimmer Biomet), shares his main insights from the cAI Conference 2024.
#CausalSky
New episode today: vist.ly/3mpfg6f
Best of $100k Causal Discovery Challenge The last weeks of November were exciting for the causal discovery community. After 12 weeks of active competition, @ADIAlab's and @crunchdao 's $100k causal discovery challenge... 1/🧵 #CausalSky
Best of $100k Causal Discovery Challenge
The last weeks of November were exciting for the causal discovery community.
After 12 weeks of active competition, @ADIAlab's and @crunchdao 's $100k causal discovery challenge...
1/🧵
#CausalSky
Indeed. I just feel scared thinking what if in our pursuit of higher "accuracy", we focused on solving the wrong problem. 🥲
Really interesting approaches all top 10. One thing I observe is that almost all of the solutions (except the 2nd place one, I think) relied very little on causal discovery algorithms/acyclicity and treated this as a classification task. I am curious what one could take out of this observation.
This week a new journey begins. I am starting as a PostDoc at
IKIM – the Institute for Artificial Intelligence in Medicine (www.ikim.uk-essen.de) where I will work at the intersection of Causal discovery and cancer research.
Looking forward to the new challenge. :-)
Well...Hello World. :-)
