Rebel Cortex: Temporal Learning In the Tree of Knowledge, Part III

Part I, II, III

A Better Way to Learn Sequences of Patterns

In my last post, I described a mechanism for learning sequences of patterns. Well, I tried it in Rebel Speech and it does not work very well. Essentially, it's too slow and it misses certain features. It turns out that there is a much faster and more effective method based on the observation that sequences are more like loops than open-ended sequences. That is to say, they repeat. We can deduce from this that all sensory signals within a sequence of patterns, whether concurrent or sequential, have the same frequency. I've written about this before in my series on Intelligent Computer Chess. Somehow I thought that there might be a better way but I should have stuck to my original instinct. In a way, this is not all that surprising as we all know from experience that repetition is the best way to remember things.

Capital

At any rate, I was going to post a detailed description of my new algorithm but I changed my mind. Yes, it's about money. I need funds to continue my research and the only way I can do it is to capitalize on what I've already discovered. I could be wrong but I am pretty confident that my Rebel Speech technology has a good chance of revolutionizing the field of speech recognition. This can go a long way toward providing me with the funds that I need. Don't worry though. I will make a demo available as soon as it's ready. Only, don't expect a full source code disclosure at this time. Sorry.