Re:Evolutionary algorithm outperforms deep-learning machines...
by Drdrdrq on Sunday July 22, 2018 @02:49AM
(#17585569)
It would be interesting to see how good the combination of EA an NN would be... This is basically us, humans: evolution + learning. Have there been any attempts to combine the two?
[ Parent
| on HN ]
Re:Evolutionary algorithm outperforms deep-learning machines...
by wmwragg on Sunday July 22, 2018 @03:09AM
(#17585622)
There have indeed, see https://en.m.wikipedia.org/wiki/Neuroevolution
Re:Evolutionary algorithm outperforms deep-learning machines...
by ufo on Sunday July 22, 2018 @05:07AM
(#17585938)
Despite the catchy names, Genetic Algorithms and Neural Networks don't work quite the same way as their biological counterparts.
Re:Evolutionary algorithm outperforms deep-learning machines...
by sprt on Sunday July 22, 2018 @05:33AM
(#17586000)
I thought GAs did, but I do lack knowledge in the area. Could you highlight some differences?
Re:Evolutionary algorithm outperforms deep-learning machines...
by ufo on Sunday July 22, 2018 @01:32PM
(#17587823)
Genetic Algorithms is a description for a wide range of techniques for adding some extra variety to search algorimhs. You maintain a population of candidate solutions and at each step you improve them by a combination of local search (like gradient descent or greedy neighbor search) and solution mixing (to help escape local minima). When doing this there are lots of ways to go about it. How to decide what population to keep, what local optimization to use, and how to mix the candidate optimizations.
If you look closely, all of these are not quite how biology works. In biology the population is decided by natural or artificial selection while in a GA there may be more factors at play (such as favoring a more diverse pool). The way solutions are represented is also different. In biology you have genes which behave according to the laws of genetics. In a GA, on the other hand, the solution representation and the operations to mutate and mix solutions are carefully planned by an intelligent designer.
Re:Evolutionary algorithm outperforms deep-learning machines...
by benp84 on Sunday July 22, 2018 @10:36AM
(#17586959)
"All models are wrong. Some are useful."
Re:Evolutionary algorithm outperforms deep-learning machines...
by flopska on Sunday July 22, 2018 @04:43AM
(#17585880)
There are some papers from Prof. Braun describing experiments where he uses mutations from genetic / evolutionary algorithms to find the optimal structure of a neural net.
Re:Evolutionary algorithm outperforms deep-learning machines...
by mrfusion on Sunday July 22, 2018 @11:01AM
(#17587079)
In my nn class Thomas tried using an ea to evolve the weights for a nn instead of using backpropagation. Surprisingly it performed way worse.
Re:Evolutionary algorithm outperforms deep-learning machines...
by nothis on Sunday July 22, 2018 @03:54AM
(#17585746)
For what it’s worth, the final paragraph says so, I believe. No links, though.
Re:Evolutionary algorithm outperforms deep-learning machines...
by awb on Sunday July 22, 2018 @03:09AM
(#17585623)
Was just thinking the same thing: Deep Evolution
Re:Evolutionary algorithm outperforms deep-learning machines...
by EGreg on Sunday July 22, 2018 @03:42AM
(#17585706)
Isn’t that what AlphaZero did with Monte Carlo Tree Search?
It basically looked only at outcomes and backtracked in a Monte Carlo tree and assigned weights without preconceived training sets.
And it outperformed the bespoke programs with existing training data.