Neural Net Talk Illustrates Business of Ideas

A simple feed-forward neural network. Here the world provides some activation to the input layer, which propagates forward through the hidden layer and eventually the output layer.

---

Mired in grading, I took a brief jaunt to Starbucks with doctoral student Wendy Maxian. This was our final meeting for our independent study this session.

We withstood the 62 degree air to walk off-campus for an iced tea. Among other things, we've been covering some philosophy of science readings. My intellectual heart was warmed when Wendy got all of the key points and made the key connections.

Then we talked about a recent lecture on campus given by parallel distributed processing czar James McClelland. I had hoped to go to the lecture with Wendy, but then the flu happened.

At any rate, today's discussion turned to the lecture and neural networks. This is a passionate topic for me, and I have not had an intense neural network discussion since moving to Lubbock.

That changed today. Out came the circles and lines, as my Indiana friends used to call them. Neural networks are beyond the scope of this post, but once you understand them, surely you will love them. It's a paradigm thing. Connectionism. Ahhh. Read here.