Re: missing thread
- Posted by useless Oct 25, 2011
- 35875 views
Nonsense, i'd happily discuss building (or growing, raising, training, etc) intelligent human-interactive computers with Greg, but i don't think this is an area of interest for him.
useless
Actually... I am taking the free Introduction to Machine Learning course from Stanford right now. I find the concepts quite intriguing and was hoping to implement some of the programming in Euphoria once I'm done with the course. Stand by for a few more weeks and we'll see what comes up! 
-Greg
I am in the middle of relocating to another state, and have only one computer with sound still enabled on it. I hope the course is still available in a few months, when i have some time and can sit back and enjoy it. There's no guarantee i'd do things the same way as the course, of course, since so many things have been tried in Ai, and so many things have not succeeded. I still think Euphoria can take off as an Ai language once it gets [drum roll] variable execution:
object command.something command.something = {insert an Eu program here} later... command.something -- the code executes
I strongly believe this is the only way an Ai can generate it's own code (make up it's own mind in novel situation) and execute in the same environment as the calling code. Tiggr does this in Mirc (build a variable and then execute it), and can dynamically load and unload whole blocks of code using Mirc's native commands to do so. Knowing if a command/keyword or variable name exists is easy in Mirc. Sadly, Mirc is quite slow, and has a few other problems.
How does the course advocate doing this? How would you do it?
useless
