in today’s entry I was free-noting abut neural networks:
Sunday, February 10, 2008 11:13 PM
Today, I took a nap that ended up being 7 hours long. All I had was earplugs, no PingPong Balls this time.
Anyway, for this nap, I ended up having a bunch of dreams. So many that I couldn’t remember specifics as I woke up. I mean, it’s like I had them, then POOF, they’re gone.
So, I started re-reading stuff on Stephen Thaler’s “Creativity Machines”. and I started wondering if perhaps these could be used somehow in conjunction with NLP. Since the creativity machines are basically just a “brainstorming” neural net, and a “training” neural net, it might be possible to program the “Training” neural net with NLP language patterns, along with corresponding situations when each pattern would be matched, like program it with the meta model language patterns and the types of statements someone might make during an interview session. In a sense, while the “learning” neural net would be learning the linguistic patterns of the Meta Model, it would also be originating new patterns which might be just as effective, if not more effective.
The possibility of combining AI with NLP was broached yesterday by Steve. So, maybe we’re onto something with this.
Another idea would be to create a different neural net for each of the NLP linguistic pattern Models. One for Meta-Model, one for Milton-Model, one for Sleight-of-Mouth, etc…
have all of those start feeding their signals into a training net and have it’s feedback basically create a neural net built up of only the effective patterns.
A new neural net that incorporates the best of all the other language models. An Artificially Intelligent Therapist?
Stumble it! | Del.Icio.us | Digg | Reddit | Subscribe to RSS Feed | Subscribe via Email