10.2 - Assumptions for Independence-Based Causal Discovery

[@BinuJasim]

Not clear why the blue path cancels the red path. A causes D so how can A is independent of D? I guess the cancellation is a possibility rather than a necessity.

[@michelspeiser5789]

In some scenarios (as in the linear example shown here) A and D are statistically independent, but not causally independent, which violates the faithfulness assumption.

[@baptistelucas8612]

the graph is actually d-separated, whether we pass by B or C

[@shubhamkumarparida5328]

A is d-separated from D when conditioned on (B,C) but A is not d separated from B unconditionally

[@wlad5351]

a secret subliminal message @0.02 ? (DEEP LEARNING VS CAUSALITY) :0 :O :o

[@moienr4104]

also at 5:05 :))