Forums before death by AOL, social media and spammers... "We can't have nice things"
|    sci.logic    |    Logic -- math, philosophy & computationa    |    262,912 messages    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
|    Message 261,291 of 262,912    |
|    Mild Shock to Kaz Kylheku    |
|    The Implicit Mechanism of Attention (sic    |
|    26 Nov 25 14:11:05    |
      XPost: sci.math, comp.theory       From: janburse@fastmail.fm              Hi              Holy cow!              How it started:              The Implicit Mechanism of Attention       It may be that a mechanism of attention       results unexpectedly and implicitly from the d       esign of the topmost, transabstractive level.       Attention would work in the following manner.       Nolarbeit Theory Journal: Part Three of Three       by Arthur T. Murray - 14 JUL 1979       https://mind.sourceforge.net/theory3.html#1979apr11              How its going:              Attention Is All You Need       The dominant sequence transduction models are       based on complex recurrent or convolutional neural       networks in an encoder-decoder configuration. The best       performing models also connect the encoder and       decoder through an attention mechanism.       Ashish Vaswani et al. 12 Jun 2017       https://arxiv.org/abs/1706.03762              Bye              Kaz Kylheku schrieb:       > Arthur Theodore Murray Obituary > "In Artificial Intelligence circles,       "Mentifex"        > (Arthur Theodore Murray) was bold and controversial.        > He composed a "theory of the mind" and developed        > AI based on his knowledge of classical languages.        > He wrote "AI4U," "AI for Latin," and other books.        > To his death, he wanted this work to be widely        > understood and useful to others."       > 1946 - 2024       > https://obituaries.seattletimes.com/obituary/arthur-murray-1089408830       >              --- SoupGate-Win32 v1.05        * Origin: you cannot sedate... all the things you hate (1:229/2)    |
[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]
(c) 1994, bbs@darkrealms.ca