Tweeted By @ml_review
A Probabilistic Representation of Deep Learning
— ML Review (@ml_review) August 29, 2019
By @UDelaware
(i) Neurons define the energy of a Gibbs distribution
(ii) the hidden layers of DNNs formulate Gibbs distributions
(iii) the whole architecture of DNNs can be interpreted as a Bayesian NNhttps://t.co/Y8Q0lE1cJC pic.twitter.com/G72qNJXZ5S