On 31/1/23 17:45, Jan Whitaker wrote: > However, I would much rather AI or heavy duty computing figure out a space > mission than someone without. Quite so... I'm not suggesting for a moment that AI may not be even better than deterministic computing in some tasks, for example initial assessment of medical diagnostic scans, or in detecting possible cases of cheating a la Turn-it-in. But in both examples the AI component only raises a signal, it doesn't make a judgement. In other words, it's a tool for use by humans.
From unreliable memory, it has been shown that a neural network can solve a class of problems which are beyond a conventional Newmann computer providing there's some form of clocking. (suggestive of our own brain waves?). I'll find more details if anyone is interested. But if it ever comes to spaceflight, I'll take a well-tested deterministic Newmann computer thanks! __Cheers, D.__ _______________________________________________ Link mailing list [email protected] https://mailman.anu.edu.au/mailman/listinfo/link
