News for hackers. Stuff that matters.

A comment on Karl Friston: a neuroscientist who might hold the key to true AI. Parent comment

Re:Karl Friston: a neuroscientist who might hold the key to ... by atschantz on Tuesday November 20, 2018 @01:35PM (#18496351)
The idea is that you learn a model by calculating the derivative of free energy with respect to your model parameters.
[ Parent | on HN ]
Re:Karl Friston: a neuroscientist who might hold the key to ... by paraschopra on Tuesday November 20, 2018 @01:38PM (#18496391)
Yes, but you have to specify a generative model (or at least put boundaries to it). Then you learn params of that model.

I was talking about learning the model structure also.

[ Parent | on HN ]
Re:Karl Friston: a neuroscientist who might hold the key to ... by atschantz on Tuesday November 20, 2018 @01:45PM (#18496465)
Some attempts have been made in the form of Bayesian model reduction [1].

The idea is to 'carve' out the structure of your model using free energy minimization.

[1] https://arxiv.org/abs/1805.07092

[ Parent | on HN ]