3 ms·
The idea is that you learn a model by calculating the derivative of free energy with respect to your model parameters.
by atschantz 8y ago
The idea is that you learn a model by calculating the derivative of free energy with respect to your model parameters.
- paraschopra 8y agoYes, 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.
- atschantz 8y agoSome 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 https://arxiv.org/abs/1805.07092