Currently Reading
I am currently working through Judea Pearl’s masterpiece on causal inference. As a statistician, this is fundamental reading that challenges many of the traditional “correlation is not causation” dogmas we were taught.
Initial Thoughts
The Ladder of Causation
Pearl introduces three levels of cognitive capability regarding cause and effect:
- Association (Seeing): What if I see X?
- Intervention (Doing): What if I do X?
- Counterfactuals (Imagining): What if I had done X?
It’s fascinating to see how most modern AI is still stuck on the first rung, while humans navigate the third rung naturally.
The Role of Models
The book argues that data alone is never enough for causal inference; we need a mental or mathematical model of the process that generated the data.
Why I’m Reading This
I want to bridge the gap between predictive ML models and causal understanding, especially for decision-making systems.
More comprehensive notes will be added once I finish the book!