Sujet : Re: Remember "Bit-Slice" Chips ?
De : 186283 (at) *nospam* ud0s4.net (186282@ud0s4.net)
Groupes : comp.os.linux.miscDate : 15. Dec 2024, 05:57:52
Autres entêtes
Organisation : wokiesux
Message-ID : <JUKdnSwFa9hH_MP6nZ2dnZfqn_qdnZ2d@earthlink.com>
References : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
User-Agent : Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.13.0
On 12/14/24 1:14 AM, rbowman wrote:
On Fri, 13 Dec 2024 22:20:45 -0500, 186282@ud0s4.net wrote:
There's also something about 92% accuracy. Eh ?
We want 100% accuracy 100% of the time. Wanna fly on a plane
structurally calculated with 92%
accuracy ?
Considering neural networks tend to be stochastic they should work well
together :)
"Stochastic" basically means "guessing".
Using such methods, a few times, might be OK for
"getting close".
But there ARE applications where "seems close enough"
is NOT good enough. Planes, spacecraft, bridges, huge
buildings, medical implants - GOTTA refine with the
hard-core/hard-math tools.
I'd suggest a TV series entitled "Engineering Disasters".
Sometimes it's the stupidest mis-calc or oversight that
leads to big flaming news stories ...