Follow
How much more likely is ¬correlation -> ¬causation than correlation -> causation?
"just generate all graphs brah it's no biggie"
Oops erdös renyo isnt dag
How much more likely is ¬correlation -> ¬causation than correlation -> causation?
"just generate all graphs brah it's no biggie"
Oops erdös renyo isnt dag
You can make a MC analysis with Erdös-Renyi directed graphs and linear (polynomial? (linear + reLU?)) functions over this, pick two random nodes, correlation, causation, bam