<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
</head>
<body>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Dear Barrack .. I fully agree with your statements below, but would add to it the following: </div>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
There is no question that the human nervous system has the ability to represent a simple symbol that is associated with a complex concept such as <cat>. We can use the written word ‘cat' or spoken word "chat” or alternatively an iconic image of a cat's face
as a symbol to convey a message about a <cat>, so our brains do learn representations for such. But this distributed representation came into being primarily for communication purposes
<b>forced by</b> and <b>helping to form</b> the social nature of our existence with each other. Later in human development (as well as in several other animals), such symbolic representations (symreps) were found to be useful as an additional constraint on
learning and reasoning about their more complex related concepts - which have a much more complex distributed representation (conreps) within the brain. </div>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif;"><span style="font-size: 16px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">In our 2023 paper (</span><span style="font-size: 16px; color: rgb(0, 120, 215); background-color: rgb(255, 255, 255);"><a href="https://arxiv.org/abs/2304.13626" data-outlook-id="c5b92b38-767d-44de-bbf4-03594f7c5bf2" style="color: rgb(0, 120, 215); text-align: left;">https://arxiv.org/abs/2304.13626</a></span><span style="font-size: 16px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">)
Tom Mitchell and I present a </span><span style="font-size: 12pt; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">Neural-Symbolic Hypothesis: "Symbols are critical to intelligence NOT because they are the building blocks of thought, but because
they are characterizations of thought that (1) allow us to explain our subsymbolic thinking to ourselves and others and (2) act as constraints on inference and learning about the world. Symbols explain our thinking and aid our thinking, but are not the foundation
of our thinking."</span></div>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div dir="ltr" class="p1" style="text-align: left; text-indent: 0px; line-height: normal; margin: 0px; font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
We conjecture that internal agent “self-communication” using symreps meant for human-to-human communication, became key to human intelligence because: (1) it provides a second, more abstract level of representation and reasoning which can occur in parallel
with subsymbolic reasoning, and (2) it places an additional constraint on learning where prior learning act as an inductive bias for learning new symbols and concepts. Shared symbols allow us to explain and justify, internally as well as externally, our decisions
and actions. And what we learn is shaped and constrained by the “lexicon” of what we recognize as symbols.</div>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div dir="ltr" class="p1" style="text-align: left; text-indent: 0px; line-height: normal; margin: 0px; font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
For the readers digest version of the paper see the Neural-Symbolic Conference abstract at
<a href="https://ceur-ws.org/Vol-3432/paper40.pdf" data-outlook-id="f59cd48f-d10e-4ae5-8c35-345985be0625">
https://ceur-ws.org/Vol-3432/paper40.pdf</a> </div>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
Best regards ... Danny </div>
<div dir="ltr" style="font-family: Aptos, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div id="ms-outlook-mobile-signature">
<p style="margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;"><span style="color: black;">==========================</span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: red;"><b>Daniel L. Silver </b></span><span style="color: black;">PhD, CIM</span></p>
<p style="margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;"><span style="color: black;">Professor Emeritus, Jodrey School of Computer Science</span></p>
<p style="margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;"><span style="color: black;">Data Scientist, Acadia Institute for Data Analytics</span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: black;">Acadia University,</span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: black;">Office 314, Carnegie Hall, </span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: black;">Wolfville, Nova Scotia Canada B4P 2R6 </span></p>
<p style="margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;"><span style="color: black;">Cell: (902) 679-9315 </span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: black;"> </span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: red;"><u><a href="http://www2.acadiau.ca/index.php" data-outlook-id="a456e230-5862-46b2-9192-3265380a3450" style="color: red; margin-top: 0px; margin-bottom: 0px;">acadiau.ca</a></u></span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: black;"><u><a href="https://www.facebook.com/acadiauniversity" title="https://www.facebook.com/acadiauniversity" data-outlook-id="8953ce14-77b3-4c05-bb87-a6f161bcba56" style="color: black; margin-top: 0px; margin-bottom: 0px;">Facebook</a></u>
<u><a href="https://twitter.com/acadiau" title="https://twitter.com/acadiau" data-outlook-id="ce2ab10d-6e4a-4a83-aed2-419e1b4fbd85" style="color: black; margin-top: 0px; margin-bottom: 0px;">Twitter</a></u>
<u><a href="https://www.youtube.com/user/AcadiaWebmaster" data-outlook-id="1d57cd0f-121a-47ea-8f38-113ab9d1d507" style="color: black; margin-top: 0px; margin-bottom: 0px;">YouTube</a></u>
<u><a href="https://www.linkedin.com/company/acadia-university?trk=biz-companies-cym" data-outlook-id="7076add0-f7e6-44f4-b0ae-4ec2ef5eb8d2" style="color: black; margin-top: 0px; margin-bottom: 0px;">LinkedIn</a></u>
<u><a href="https://www.flickr.com/photos/acadiauniversity/albums" title="https://www.flickr.com/photos/acadiauniversity/albums" data-outlook-id="abfe3842-5ffe-42ab-9ba0-3de06fb1e3c9" style="color: black; margin-top: 0px; margin-bottom: 0px;">Flickr</a></u></span></p>
<p style="text-align: left; margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<span style="color: black;"><img src="cid:image001.png@01D7A706.136138E0" alt="id:image001.png@01D366AF.7F868A70" id="Picture_x0020_1" width="270" height="55" style="width: 2.8229in; height: 0.5833in; margin-top: 0px; margin-bottom: 0px;"></span></p>
<p style="margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;"> </p>
<p style="margin: 0cm;"><span style="font-family: Calibri, sans-serif; font-size: 11pt;"> </span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0);"><b>From:
</b></span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">Connectionists <connectionists-bounces@mailman.srv.cs.cmu.edu> on behalf of Barak A. Pearlmutter <barak@pearlmutter.net></span><span style="font-family: Calibri, sans-serif; font-size: 11pt; color: rgb(0, 0, 0);"><br>
</span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0);"><b>Date:
</b></span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">Tuesday, August 25, 2026 at 12:25 PM</span><span style="font-family: Calibri, sans-serif; font-size: 11pt; color: rgb(0, 0, 0);"><br>
</span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0);"><b>To:
</b></span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">connectionists@mailman.srv.cs.cmu.edu <connectionists@mailman.srv.cs.cmu.edu></span><span style="font-family: Calibri, sans-serif; font-size: 11pt; color: rgb(0, 0, 0);"><br>
</span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0);"><b>Subject:
</b></span><span style="font-family: Aptos; font-size: 16px; color: rgb(0, 0, 0); background-color: rgb(255, 255, 255);">Re: Connectionists: AI@50 videos...</span></p>
<div dir="ltr" style="margin: 0cm; font-family: Calibri, sans-serif; font-size: 11pt;">
<br>
</div>
</div>
<div id="mail-editor-reference-message-container">
<div class="PlainText" style="font-size: 11pt;">CAUTION: This email comes from outside Acadia. Verify the sender and use caution with any requests, links or attachments.<br>
<br>
> Plate (2002): “Another equivalent property is that in a distributed representation one cannot interpret the meaning of activity on a single neuron in isolation: the meaning of activity on any particular neuron is dependent on the activity in other neurons
(Thorpe, 1995).”<br>
<br>
> Thorpe (1995, p. 550): “With a local representation, activity in individual units can be interpreted directly … with distributed coding individual units cannot be interpreted without knowing the state of other units in the network.”<br>
<br>
> Elman (1995, p. 210): “These representations are distributed, which typically has the consequence that interpretable information cannot be obtained by examining activity of single hidden units.”<br>
<br>
People blithely toss out statements like that, but local<br>
uninterpretability does not follow from the definition of a<br>
distributed representation.<br>
<br>
It is very easy to construct distributed representations whose<br>
components *can* be understood in isolation (the bits of a standard<br>
binary representation of a natural number ≤2³², say) or whose<br>
components cannot be understood in isolation (a code for a natural<br>
number ≤2³² whose bits each have an associated set of half the numbers<br>
in the domain picked at random, say).<br>
<br>
The "big regions" distributed representation for points in 2D space<br>
from the PDP books is locally interpretable.<br>
<br>
Naturally, all other things being equal, given the choice we'd choose<br>
distributed representations whose parts are interpretable in<br>
isolation. Does the brain do that? Well, deep learning networks seem<br>
to, and experimental data seems consistent with that notion.<br>
<br>
--Barak Pearlmutter.<br>
<br>
</div>
</div>
</body>
</html>