search

ascavis / machine-magic

GNU Affero General Public License v3.0
0 stars 0 forks source link

2dos #2

Open baconstrudel opened 9 years ago

baconstrudel commented 9 years ago

-figure out a concept for how the data is preprocessed automatically -settle for a couple of different Networks to try -implement Networks -Magic the data into beautifully categorized asteroids

seyr commented 9 years ago

I take it you have already decided on the algorithm (class)? Neural networks? Can you give me the name/overview link maybe so I can get up to speed?

baconstrudel commented 9 years ago

ok so the inital Approach would be a deep belief network with dropout and pretraining

baconstrudel commented 9 years ago

second approach would be a 2 step approach, the first step being a dimension reduction with a sanger-net or kohonen-map and then a perceptron for classification, maybe a two layer-network with RBM and perceptron

baconstrudel commented 9 years ago

rbm = Restricted Boltzmann machine

baconstrudel commented 9 years ago

I meant neural gas not sanger-net

baconstrudel commented 9 years ago

how do you get along?

seyr commented 9 years ago

Still reading. Did you already start with code?

baconstrudel commented 9 years ago

I think we might want to do it together

seyr commented 9 years ago

I see. You had libraries yesterday, no? Did you get my messages? (dragon.fly@jabber.org)

seyr commented 9 years ago

(have to get some food now, I'll be back)