Home

Blogs

Forums

News

Portfolios

Projects

Tutorials

21 users logged in

New? Learn about game development.

Before posting, review our community guidelines.

Chat in the GameDev.net Discord!

🎉 Celebrating 25 Years of GameDev.net! 🎉

Not many can claim 25 years on the Internet! Join us in celebrating this milestone. Learn more about our history, and thank you for being a part of our community!

Back to Artificial Intelligence

The adventures of Mr Hopfield and his energy landscape

Mr Nonsense · 2002-10-14T07:39:44

I''m having problems with a simple hopfield network. Basically, I''m storing 3 patterns in a network of 63 neurons. The network will converge to patterns 1 and 2, but will never converge to pattern 3. Even if I set the initial state of the network identical to pattern 3, the network will always converge to pattern 2. I''m aware of the problems of storage capacity in a hopfield network. Using certain formulas you can approximate that a hopfield network with 63 neurons should be able to hold and recall upto 10 patterns. So the problem is, why is my network not able to hold and recall more than 2 patterns? I''m trying to visualise the energy landscape of the network to try and understand why it doesn''t work. It''s my understanding that local minima on the energy surface represent stored patterns. Since a hopfield network will reduce it''s energy on every iteration, it should get stuck in a local minimum and thus converge on a stored pattern. So why is it that when the network''s initial state is the same as a stored pattern it converges on another pattern? I''m assuming the reason has something to do with how similar the patterns are. Is it possible that if two patterns are similar then the local minimum of one pattern can erase the local minimum of another pattern? Does anybody have any ideas? Can anybody help me? Have I finally lost it? Thanks

Artificial Intelligence Programming

Started by Mr Nonsense October 04, 2002 07:05 AM

13 comments, last by Mr Nonsense 21 years, 9 months ago

Mr Nonsense

Author

126

October 08, 2002 03:06 AM

I might live in a house up a tree, and sleep in a boat, but that''s just plain nonsense.

Timkin

864

October 08, 2002 09:20 PM

Rolling On Floor Laughing My Ass Off For A Very Long Time

fup

463

October 09, 2002 01:11 AM

ah!

ai-junkie.com

My Website: ai-junkie.com | My Books: 'Programming Game AI by Example' & 'AI Techniques for Game Programming'

edotorpedo

198

October 14, 2002 07:12 AM

BTW: the book I was refering to is:

"The essence of Neural Networks'' by Robert Callan. I implemented the experiment exactly the way it is described on page 89.

Damn that Robert!

Good book for all other networks though.

Edo

Edo

Mr Nonsense

Author

126

October 14, 2002 07:39 AM

That was the book I used. It''s a nice introductory book to neural networks, although it does have one or two serious misprints (especially in the backpropagation algorithm).

Just in case nobody believes my claims about a hopfield network, I found a little java applet. See how many patterns you can store before the network goes arse over tit

http://www.psych.uchicago.edu/~regier/hopfield/

🎉 Celebrating 25 Years of GameDev.net! 🎉

The adventures of Mr Hopfield and his energy landscape

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

🎉 Celebrating 25 Years of GameDev.net! 🎉

The adventures of Mr Hopfield and his energy landscape

This topic is closed to new replies.

Popular Topics

Recommended Tutorials

Reticulating splines