• Welcome to BellGab.com Archive.
 

Weka Data Mining/Machine Learning

Started by Mind Flayer Monk, June 02, 2014, 07:16:27 PM

A nice Data Mining/Machine Learning software program if anyone is interested in learning.
Weka is fairly light and has far less overhead than Mahout.
A link to the main page
http://www.cs.waikato.ac.nz/ml/weka/index.html

and to a Coursera class you can use to train
https://weka.waikato.ac.nz/dataminingwithweka

McPhallus

Quote from: Mind Flayer Monk on June 02, 2014, 07:16:27 PM
A nice Data Mining/Machine Learning software program if anyone is interested in learning.
Weka is fairly light and has far less overhead than Mahout.
A link to the main page
http://www.cs.waikato.ac.nz/ml/weka/index.html

and to a Coursera class you can use to train
https://weka.waikato.ac.nz/dataminingwithweka

;D

area51drone

Very interesting Monk, I definitely want to look into this sometime soon.


Eugene Goostman, a computer programme pretending to be a young Ukrainian boy, successfully duped enough humans to pass the iconic test.

Fooled the experts.

area51drone

I'd love to know the "experts" questions, as I didn't feel "what did you eat for dinner?" to be that off the wall.   They list a couple of the experts in the article, and they're no experts, they're just random semi famous people.

Not impressed.

I wish I didn't have to work, and could just focus on AI.  I think the real trick is going to be to create a digital copy of a human brain, and just let the computer go to work with its various stimuli.  There are a few projects out there attempting to do this, but the human brain has so many neurons and connections it's nearly impossible to with our current technology according to what I've read.   Wiki says 15-30 billion neurons.   That doesn't seem like that many.

In other AI news, it sounds like Watson is now being commercially used in hospitals starting this January, and the size of the machine has gone down to "three pizza boxes."   Interesting stuff!  http://en.wikipedia.org/wiki/IBM_Watson

Quote from: area51drone on June 10, 2014, 04:27:15 AM
I think the real trick is going to be to create a digital copy of a human brain, and just let the computer go to work with its various stimuli.  There are a few projects out there attempting to do this, but the human brain has so many neurons and connections it's nearly impossible to with our current technology according to what I've read.   Wiki says 15-30 billion neurons.   That doesn't seem like that many.


I think there was one where they were trying to clone and upload simpler brains-like a lobster's or a nematode's brain.



area51drone

Quote from: Mind Flayer Monk on June 10, 2014, 12:34:03 PM
I think there was one where they were trying to clone and upload simpler brains-like a lobster's or a nematode's brain.

Yeah they are, and even putting them on chips and FPGA's so they can run quicker.   Still no major breakthroughs though.  So far, Watson is the best AI I've seen, and that isn't saying much considering all it can do is answer questions about data sets.

bigchucka

Quote from: area51drone on June 10, 2014, 02:20:44 PM
Yeah they are, and even putting them on chips and FPGA's so they can run quicker.   Still no major breakthroughs though.  So far, Watson is the best AI I've seen, and that isn't saying much considering all it can do is answer questions about data sets.

I'll take data sets for 1000, Alex.

Powered by SMFPacks Menu Editor Mod