AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Is It Possible to Learn the Causal Structure?
There's a lot of discussion at the moment about causal representation, learning from observationar data. In geometric deep learning, we've hand baked these inductive priers, and they work very well. Do you think we could do the same thing for causal structure? Or do you think it's actually empirically learnable from observational data? If you just have observational data, ther you can actually show that they are a circumstance in which you aly callate that.