News for hackers. Stuff that matters.

A comment on Exploring Weight Agnostic Neural Networks. Parent comment

Re:Exploring Weight Agnostic Neural Networks by patresh on Thursday August 29, 2019 @08:31AM (#20828873)
Each architecture is tested multiple times against different samples of the shared weight

From the paper :

(1) An initial population of minimal neural network topologies is created

(2) each network is evaluated over multiple rollouts, with a different shared weight value assigned at each rollout

(3) networks are ranked according to their performance and complexity

(4) a new population is created by varying the highest ranked network topologies, chosen probabilistically through tournament selection

[ Parent | on HN ]
Re:Exploring Weight Agnostic Neural Networks by TekMol on Friday August 30, 2019 @02:49AM (#20836427)
What is a "shared weight"?
[ Parent | on HN ]
Re:Exploring Weight Agnostic Neural Networks by patresh on Saturday August 31, 2019 @05:48AM (#20845443)
Usually in neural networks, each neuron has their own weights with different values that will be tuned during training.

Shared weight here means that every neuron in the network shares the exact same weight value.

[ Parent | on HN ]