Content deleted Content added
→More generality: new section |
''create page'': decoding megacluster |
||
Line 34:
Why is this article only discussing binary codes? Decoding (and code theory in general) works fine over <math>\mathbb{F}_q^n</math>. [[User:Xpple|Xpple]] ([[User talk:Xpple|talk]]) 21:50, 21 February 2023 (UTC)
== ''create page'': [[decoding megacluster]] (AI and ANNs) ==
Create ANN (artificial neural network) clusters = ANNs of different design each (Brodmann-like area) used to create a big brain.
Use typical ANNs, Flexible ANNs, non-ANN programming and generative pre-trained transformer-like (similar idea; not necessarily an exact copy).
Combine all the above to create a Flexible Artificial Neural Network Cluster (complete brain) for decoding. It requires coherence centers [to test and edit the outputs]; otherwise it doesn't work as a single cluster/ brain.
Run it on a supercomputer. The big brain cluster must clone itself inside the supercomputer and test alternative approaches. If it gets decoding hints, the system is supposed to share it inside the system; but with some limitation, because a false hint shouldn't destroy the overall effort.
| |||