<div dir="ltr"><div dir="ltr"><div>Andreas Wichert has mentioned Dave Touretzky's 1998 query to the connectionists mailing list: "Is connectionist symbol processing dead?"</div><div><br></div><div>The responses to that query were collected and published in Neural Computing Surveys, which doesn't exist anymore.<br></div><div>However, the paper is still available from the WaybackMachine:<br></div><div><a href="https://web.archive.org/web/20170706013814/ftp://ftp.icsi.berkeley.edu/pub/ai/jagota/vol2_1.pdf">https://web.archive.org/web/20170706013814/ftp://ftp.icsi.berkeley.edu/pub/ai/jagota/vol2_1.pdf</a></div><div><br></div><div>Cheers</div><div><br></div><div>Ross<br></div><div><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, 21 Jul 2022 at 20:31, Andrzej Wichert <<a href="mailto:andreas.wichert@tecnico.ulisboa.pt">andreas.wichert@tecnico.ulisboa.pt</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div style="overflow-wrap: break-word;">Dear Juergen,<div><br></div><div><div><div style="margin:0px;font-stretch:normal;line-height:normal">Symbols do not, by themselves, represent any utilizable knowledge, they cannot be used for a definition of similarity criteria between themselves. The use of symbols in algorithms which imitate human intelligent behavior led to the famous physical symbol system hypothesis  \index{physical symbol system hypothesis} by Newell and Simon (1976)   ``The necessary and sufficient condition for a physical system to exhibit intelligence is that it be a physical symbol system.''  Symbols  are not present in the world; they are the constructs of a human mind and simplify the process of representation used in communication and problem solving.  </div></div><div><br></div><div>A Turing machine can simulate any algorithm, the same with RNN, but they do not explain the human problem solving.</div><div><br></div><div>There is a difference, for example human problem solving can be described by production systems. The most successful model is the SOAR architecture. Such systems were very successful, they learn and can give you an explanation for their doing.</div><div><br></div><div><div style="margin:0px;font-stretch:normal;line-height:normal">in August 1998 Dave Touretzky asked on the connectionistic e-mailing list:</div><div style="margin:0px;font-stretch:normal;line-height:normal">``Is connectionist symbol processing dead?’'</div><div style="margin:0px;font-stretch:normal;line-height:normal"><br></div><div style="margin:0px;font-stretch:normal;line-height:normal"><br></div><div style="margin:0px;font-stretch:normal;line-height:normal">From <a href="mailto:ml-connectionists-request$@$mlist-1.sp.cs.cmu.edu" target="_blank">ml-connectionists-request$@$mlist-1.sp.cs.cmu.edu</a>  Tue Aug 11 17:35:10 1998</div><div style="margin:0px;font-stretch:normal;line-height:normal">From: Dave$\_$<a href="mailto:Touretzky$@$cs.cmu.edu" target="_blank">Touretzky$@$cs.cmu.edu</a></div><div style="margin:0px;font-stretch:normal;line-height:normal">To: <a href="mailto:connectionists$@$cs.cmu.edu" target="_blank">connectionists$@$cs.cmu.edu</a></div><div style="margin:0px;font-stretch:normal;line-height:normal">Subject: Connectionist symbol processing: any progress?</div><div style="margin:0px;font-stretch:normal;line-height:normal">Date: Tue, 11 Aug 1998 $03:34:27$ -0400</div><div style="margin:0px;font-stretch:normal;line-height:normal"><br></div><div style="margin:0px;font-stretch:normal;line-height:normal">I'd like to start a debate on the current state of connectionist symbol</div><div style="margin:0px;font-stretch:normal;line-height:normal">processing?  Is it dead?  Or does progress continue?  ...  People had gotten</div><div style="margin:0px;font-stretch:normal;line-height:normal">some interesting effects with localist networks, by doing spreading activation</div><div style="margin:0px;font-stretch:normal;line-height:normal">and a simple form of constraint satisfaction....  This approach does not create</div><div style="margin:0px;font-stretch:normal;line-height:normal">new structure on the fly, or deal with structured representations or variable</div><div style="margin:0px;font-stretch:normal;line-height:normal">binding.  Those localist networks that did attempt to implement variable binding</div><div style="margin:0px;font-stretch:normal;line-height:normal">did so in a discrete, symbolic way that did not advance the parallel constraint</div><div style="margin:0px;font-stretch:normal;line-height:normal">satisfaction/heuristic reasoning agenda of earlier spreading activation</div><div style="margin:0px;font-stretch:normal;line-height:normal">research.  ...  So I concluded that connectionist symbol processing had reached</div><div style="margin:0px;font-stretch:normal;line-height:normal">a plateau, and further progress would have to await some revolutionary new</div><div style="margin:0px;font-stretch:normal;line-height:normal">insight about representations.  ...  The problems of structured representations</div><div style="margin:0px;font-stretch:normal;line-height:normal">and variable binding have remained unsolved.  No one is trying to build</div><div style="margin:0px;font-stretch:normal;line-height:normal">distributed connectionist reasoning systems any more, like the connectionist</div><div style="margin:0px;font-stretch:normal;line-height:normal">production system I built with Geoff Hinton...</div><div style="margin:0px;font-stretch:normal;line-height:normal"><br></div></div><div style="margin:0px;font-stretch:normal;line-height:normal"><br></div><div>24 years past and not much progress was done. It seems that the progress is only related to pure brute force of computers, but not much insight beside a wishful thinking. The whole DL movements stops the progress in understanding how the brain works. We need some new fresh ideas beside error minimization.</div><div><br></div><div>I think one of the main problems is the publish or perish altitude, the famous Google impact factor. One does not care what some one is doing, one just checks his Google impact factor.. This is like the saying, eat more shit, one million flies cannot be wrong. </div><div>Like some mathematician said, computer science is not science at all, but it force us to follow its ideas.</div><div><br></div><div>Andreas</div><div><br></div><div><br></div><div>
<br><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none">--------------------------------------------------------------------------------------------------</span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none">Prof. Auxiliar Andreas Wichert   </span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none"><a href="http://web.tecnico.ulisboa.pt/andreas.wichert/" target="_blank">http://web.tecnico.ulisboa.pt/andreas.wichert/</a></span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none">-</span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none"><a href="https://www.amazon.com/author/andreaswichert" target="_blank">https://www.amazon.com/author/andreaswichert</a></span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none">Instituto Superior Técnico - Universidade de Lisboa</span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none">Campus IST-Taguspark </span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none">Avenida Professor Cavaco Silva                 Phone: +351  214233231</span><br style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none"><span style="color:rgb(0,0,0);font-family:Helvetica;font-size:12px;font-style:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;text-decoration:none;display:inline;float:none">2744-016 Porto Salvo, Portugal</span>
</div>
<br></div></div></blockquote></div></div>