On: Artificial Intelligence with Erlang: The Domain of...

[cjauvin]

I'm curious, why do you say it's unfortunate? (Don't get me wrong I am also of that opinion, more or less, so I'm just curious about yours)

[vokolk → cjauvin]

Check out this paper by Turin-award winner Judea Pearl for fundamental limitations of current approaches to ML: https://arxiv.org/abs/1801.04016

[fjsolwmv → vokolk]

I think that argument ignores that neural nets can learn the models that Pearl believes are important. And it's quite plausible that human reasoning works in similar way.

Pearl's career is all about causal networks, so he's slightly biased when it comes to survey like this.

[vokolk → fjsolwmv]

So you are saying that Pearl's argument that current ML system cannot reason about interventions and retrospection (as he defines them) are wrong?