Population-based training of neural networks

Re:Population-based training of neural networks by deepnotderp on Monday November 27, 2017 @10:47AM (#15788176)

So basically SGD based NEAT?

[ Parent | on HN ]

Re:Population-based training of neural networks by Rhapso on Monday November 27, 2017 @01:14PM (#15789577)

NEAT paper link I post a lot on HN: http://nn.cs.utexas.edu/downloads/papers/stanley.ec02.pdf

Looks like like a "not a GA we promise" on hyper-parameter selection (which is cool if unnecessarily obtuse)

[ Parent | on HN ]

Re:Population-based training of neural networks by YaxelPerez on Monday November 27, 2017 @11:51AM (#15788720)

NEAT just uses GA to generate a network topology and weights. From what I read, I think this is just a fancy way to parallelize searching for optimal hyperparameters.

[ Parent | on HN ]

Re:Population-based training of neural networks by deepnotderp on Monday November 27, 2017 @11:55AM (#15788754)

Well, okay, what I meant is that it's literally applying evolutionary algorithms for hyperparameter optimization. Calling it some souped up BS like "population based training" seems like Deepmind marketing is getting out of hand...

[ Parent | on HN ]

Re:Population-based training of neural networks by nametube on Monday November 27, 2017 @03:40PM (#15791075)

Evolutionary Algorithms are a subset of Population Based heuristics. I don't think its "souped up BS" to use the term.

[ Parent | on HN ]

Re:Population-based training of neural networks by kirillseva on Monday November 27, 2017 @03:20PM (#15790850)

Maybe they have a quota on how many impressions they need to get per quarter?

[ Parent | on HN ]