Connectionists: Ensemble computing

Mon Jun 17 03:18:42 EDT 2024

Dear Asim:

Place cells are spatio-temporal constructs subject to phase coding. They definitely fall under the category of population code. Single place cells are imprecise in defining the position, while the spatio-temporal voting of the neuronal ensemble gives better results. In addition, the spatial response of a single neuron can't be seen as the abstract representation of 2D (3D for bats) geometry. There are other networks that respond too and are in phase synchrony (phase locking – DOI: 10.1126/sciadv.abm6081<https://doi.org/10.1126/sciadv.abm6081>) offering additional metric and topology plus geometrical information for the forecasting capabilities of place cells enhancing further the relevance and the meaning of population coding.

Furthermore, the existence of the representation of 2D (3D) geometry may not involve conscious thinking in terms of the belonging concepts. They may grant the precise decision about just-in-time action and nothing else. Note that conscious knowledge of actions in humans are delayed by about 200 msec relative to the neuronal signals launched by the motor cortex to initiate those actions.

In turn,  “meaning” does not seem to fly in the face of population coding.

Best,

Andras Lorincz
Fellow of the European Association for Artificial Intelligence
Department of Artificial Intelligence
Faculty of Informatics
Eotvos Lorand University
Budapest, Hungary

________________________________
From: Connectionists <connectionists-bounces at mailman.srv.cs.cmu.edu> on behalf of Asim Roy <ASIM.ROY at asu.edu>
Sent: Sunday, June 16, 2024 10:58 PM
To: Weng, Juyang <weng at msu.edu>; Gabriele Scheler <gscheler at gmail.com>; connectionists at mailman.srv.cs.cmu.edu <connectionists at mailman.srv.cs.cmu.edu>
Subject: Re: Connectionists: Ensemble computing

Some have difficulty accepting findings that were rewarded with Nobel prizes. These scientists were rewarded for finding “meaning” in the activations of single neurons. That flies in the face of population coding. And, in science, you can repeat these experiments and verify these findings again and again. Some see lots of activations and think that must be population coding.

Asim Roy

Professor, Information Systems

Arizona State University

Asim Roy | ASU Search<https://search.asu.edu/profile/9973>

Lifeboat Foundation Bios: Professor Asim Roy<https://lifeboat.com/ex/bios.asim.roy>

From: Connectionists <connectionists-bounces at mailman.srv.cs.cmu.edu> On Behalf Of Weng, Juyang
Sent: Saturday, June 15, 2024 7:28 AM
To: Gabriele Scheler <gscheler at gmail.com>; connectionists at mailman.srv.cs.cmu.edu
Subject: Re: Connectionists: Ensemble computing

Dear Gabriele Scheler,

     You wrote, "Each symbol stands for one pattern".  This is an invalid statement, meaningless in mathematics and computer science.

    (1) I suggest that you review the term One-to-One correspondence required for symbolic representations in discrete mathematics and computer science (like Turing machines).

    (2) Counter Example 1: While my brain senses your face, many neurons in my brain will fire.   Do they stand for the same pattern (your face)?  This means one-to-many mapping, defeating the one-to-one correspondence in (1).

    (3) Counter Example 2: While my brain senses your face at different viewing angles, distances, and lighting conditionss, my motor neurons will fire to produce Gabriele Scheler.  Any of motor neurons stand for many patterns of your face, not only "one pattern".  This means many-to-one mapping, defeating the one-to-one correspondence in (1).

    With the above proof, "Each symbol stands for one pattern" is meaningless in mathematics and computer science.

    Best regards,

-John Weng

On Sat, Jun 15, 2024 at 9:46 PM Gabriele Scheler <gscheler at gmail.com<mailto:gscheler at gmail.com>> wrote:

Oh st simpkicissismus.

Symbolic neurons have connections with each other. Each symbol stands for one pattern. They also have connections with their feature patterns. But we compute with the symbolic neurons alone. There are no examples given for that.

So by computing with symbolic neurons we can use exact reasoning. But when we want to know the meaning of a symbol we can then turn to their features. The clever part - since you have trouble reading - is that we can use only the symbolic neurons, we do not ha e to stimulate them in a way that their feature parts become active. In that case, 1 symbol, 1 neuron.

________________________________

From: Connectionists <connectionists-bounces at mailman.srv.cs.cmu.edu<mailto:connectionists-bounces at mailman.srv.cs.cmu.edu>> on behalf of Gabriele Scheler <gscheler at gmail.com<mailto:gscheler at gmail.com>>
Sent: Monday, June 10, 2024 8:03 AM
To: connectionists at mailman.srv.cs.cmu.edu<mailto:connectionists at mailman.srv.cs.cmu.edu> <connectionists at mailman.srv.cs.cmu.edu<mailto:connectionists at mailman.srv.cs.cmu.edu>>
Subject: Connectionists: Ensemble computing

This paper may be relevant to the discussion on symbolic computing in the brain.

https://www.biorxiv.org/content/10.1101/2023.12.22.573036v2<https://urldefense.com/v3/__https:/www.biorxiv.org/content/10.1101/2023.12.22.573036v2__;!!HXCxUKc!3XvVBOS7Ewzj9hEc7LIS4jJ1KaOBIEg96-dgQQoFBKgSXwmbWIv9bbBw_q5aEBIrkXofN97mq1r_Tw$>

It leverages ideas about ensembles in the brain and outlines a basic scheme for symbolic abstraction .

Gabriele

--

Dr. Gabriele Scheler
Carl Correns Foundation for Mathematical Biology

1030 Judson Dr
Mountain View, CA 94040

https://www.theoretical-biology.org<https://urldefense.com/v3/__http:/www.theoretical-biology.org__;!!HXCxUKc!3XvVBOS7Ewzj9hEc7LIS4jJ1KaOBIEg96-dgQQoFBKgSXwmbWIv9bbBw_q5aEBIrkXofN95zer6I-w$>

Please re-send an email if I fail to respond.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20240617/0d6d65c5/attachment.html>