Posts Tagged With 'learning'
Artificial Intelligence - Part Three Of A Billion - Hidden Models / Neural Equivalence?
Computation neural networks are incredibly interesting. The principles by which they work are simple, but the problems that they could solve are extremely complex. The trouble with neural networks is that they are computationally expensive and fiendishly difficult to train. This has lead to research in other techniques and learning methods.
...
posted May 9, 2010. It has four comment(s).
tagged with: AI ai-series hidden models HMM learning neural-network neurons
Manifold Learning - Some Results
I've been researching manifold learning recently, see here, and to this end, I'm designing a generative algorithm to learn manifolds based on topological methods. It's not done yet, but I have some initial results, based on computing geodesics, that are quite interesting. I have a fairly large dataset of hand-written twos which I've fed into my program. Each line in the image below represents a geodesic.
...
posted April 12, 2010. It has five comment(s).
tagged with: AI geodesic images learning manifold results vision
