- Comp-neuro - lists.cnsorg.org

[comp-neuro] job advertisement | 1 fully funded PhD position (E13) | Application deadline: Nov.18, 2024

by Daniel Schmid

Vacancy for dissertation/PhD project *Learning Search and Decision Mechanisms in Medical Diagnosis* We are looking for a new PhD student for a fully funded interdisciplinary research project investigating the computational mechanisms underlying the perceptual processes in medical imaging. *Brief project description and goals* Medical experts routinely screen images for signs of abnormality indicative of a disease. Experimental evidence shows that these experts, based on their training, are able to make a diagnose above chance level even after screening images for only a few seconds. Such visual inspection utilizes global ensemble scene statistics, or “scene gist”, to provide contextual guidance information. The goal of the PhD project is to investigate how such global, contextual gist information can be computed and incorporated in machine vision approaches, such as deep neural networks for natural or medical image processing. In particular, a goal is to develop neural mechanisms that can be integrated into existing pre-trained CNNs to compute scene gist for rapid global decision-making and contextual guidance in medical images. Such mechanisms should explain how such global feature compositions are learned to increase the classification specificity and, at the same time, do not equip an observer with the ability to precisely localize such feature compositions indicative of the evidence. At the implementation level, computational units (neurons) in hierarchical CNN architectures will be extended by integrating local bottom-up feature extraction mechanisms with top-down modulatory contextual fields. We focus on these mechanisms to learn the integration of information streams in counter-stream networks by employing novel two-point information integration units inspired by recent findings from neuroscience. These investigations will contribute to develop novel CNN architectures by integrating feedforward and feedback data streams that operate at different spatio-temporal scales and feature types inspired by computational mechanisms in biological vision. The combined global and local information is expected to provide more detailed explanations on a mechanistic level how radiologists steer their attention and search capacities and learn to improve such skills. *Requirements* - MSc degree in Computer Science, Engineering, Physics, Mathematics, Cognitive Systems/Science or equivalent relevant background - Experience in computational vision, neural networks, machine learning; with strong mathematical/physics foundation - Experience in developing biologically inspired models – or interested in learning how to develop such mechanisms and models - Experience in analysis of large data and evaluation of results - Strong coding skills - Personal skills: working in a team, but should be able to work independently, is strongly focused and enthusiastic about the project theme, is self-motivated and takes responsibility for reaching milestones - Experience and talent in scientific writing (with proficient written & oral English) *Contact & application* The project is part of the newly founded research training group “KEMAI” that focuses on integrating knowledge and learning based approaches for better medical AI (https://kemai.uni-ulm.de/ for more information). The offered PhD position is fully funded (E13 salary) and also benefits from an interdisciplinary setting, the ability to connect with peers from related projects and a clearly structured PhD program. Application materials are expected until November 18, 2024. They should contain a CV, statement of research interest and background in relation to the KEMAI training group and project theme C2. A cover letter should inform about the expected date of availability and name of two referees. Submit your application via the KEMAI website (preferred) or to Prof. Heiko Neumann (heiko.neumann(a)uni-ulm.de) or reach out for more information. Best regards, Daniel Schmid -- Daniel Schmid, M.Sc. Institute of Neural Information Processing Faculty of Engineering, Computer Science and Psychology Ulm University James-Frank-Ring D-89081 Ulm Germany e-mail: daniel-1.schmid(a)uni-ulm.de

Oct. 28, 2024

Justin Wood speaking on November 12 in Developing Minds global online lecture series

by Jochen Triesch

Dear colleagues, On November 12, the Developing Minds global online lecture series is proud to host Justin Wood from Indiana University Bloomington, USA, speaking on: "Radical empiricism: The origins of knowledge as a mini-evolution“ Tuesday, November 12, 2024 9:00 am EST (Eastern Standard Time, US) 14:00 UTC (Universal Coordinated Time) 15:00 CET (Central European Time) 23:00 JST (Japan Standard Time) The zoom link/credentials are: https://uni-frankfurt.zoom-x.de/j/62980753356?pwd=0J6BtOaAy20YyanXX1gpkB3Ta… Meeting-ID: 629 8075 3356 Kenncode: 904159 Abstract: What are the origins of knowledge? A common theoretical strategy is to isolate core primitives of the mind, such as systems for reasoning about objects, space, number, agents, and language. The problem with this strategy is it disconnects brain development from general fitting principles that underlie chemistry, evolution, culture, and artificial intelligence. I argue that development can be understood as a “mini-evolution,” in which core mental skills are the products of generic evolution-like fitting processes. In evolution, life started from scratch, and species emerged as DNA adapted to the world via blind fitting. Likewise, in development, knowledge starts from scratch, and mental skills emerge as individual brains adapt to the world via blind fitting. Historically, the main roadblock for testing fitting theories has been the lack of benchmarks and models for measuring whether fitting models learn like brains. To directly compare learning across brains and models, both must be trained with the same data and tested on the same tasks. We propose a solution—called “Newborn Embodied Turing Tests” (NETTs)—in which newborn animals and fitting models are reared in the same environments and tested with the same tasks. I’ll describe a series of NETTs showing that generic fitting models (transformers) develop many core mental skills, including orientation selectivity, visual binding, shape-based vision, invariant object recognition, imprinting, collective behavior, and social preferences. These skills develop spontaneously when generic fitting models fit to prenatal and postnatal experiences. There is no need to postulate mysterious core primitives to explain the rapid development of domain-specific knowledge. Rather, core knowledge can be the product of generic fitting machinery: a radical empiricist view of the origins of knowledge. This mini-evolution view unifies chemistry, evolution, development, culture, and artificial intelligence under a common fitting framework, with shared general principles. Short Bio: Justin Wood is an Associate Professor of Informatics at Indiana University Bloomington. He is affiliated with programs in Animal Behavior, Neuroscience, Cognitive Science, and Psychological & Brain Sciences. He received his B.A. from University of Virginia and M.S. & Ph.D. from Harvard University. Dr. Wood has studied the psychological abilities of a range of populations, including human adults, children, infants, chimpanzees, wild monkeys, starlings, and newborn chicks. He works at the intersection of developmental psychology, neuroscience, virtual reality, and artificial intelligence to characterize the origins and computational foundations of intelligence. The talk will be recorded and made available for later viewing. For more information on the talk series and recordings of previous events, please visit: https://sites.google.com/view/developing-minds-series/home Best regards, Jochen Triesch -- Prof. Dr. Jochen Triesch Johanna Quandt Chair for Theoretical Life Sciences Frankfurt Institute for Advanced Studies and Goethe University Frankfurt http://fias.uni-frankfurt.de/~triesch/ Tel: +49 (0)69 798-47531 Fax: +49 (0)69 798-47611

Oct. 28, 2024

Re: Join Us for the Statistical Physics of Cognition Workshop – Institute of Physics, London

by Rajpal, Hardik

Dear Colleagues, We are pleased to announce that registrations are now open for the Statistical Physics of Cognition Workshop, taking place on the 25th and 26th of November 2024 at the Institute of Physics, King’s Cross, London. You can register for the event here: https://iop.eventsair.com/spc2024/register We would also like to remind everyone that the abstract submission deadline for the poster session is fast approaching on 31st October 2024. PhD students and early career researchers are particularly encouraged to submit abstracts for this exciting opportunity to share your work. Please submit your abstracts at the following link: https://iop.eventsair.com/spc2024/abstract-submission For further details about the workshop, you can explore the program here: https://iop.eventsair.com/spc2024/programme Additionally, we are thrilled to announce a public event featuring a fireside chat between Prof. John Krakauer and Prof. Karl Friston on the intriguing topic, "Can Reductionism Explain the Mind?" This discussion will take place on the evening of 25th November, followed by a drinks reception. We look forward to welcoming you to an interdisciplinary gathering that will bring together experts from neuroscience and physics to explore cognition in innovative ways. If you have any questions, please feel free to reach out, and we hope to see you there! Sincerely, Hardik Rajpal Research Associate Organizing Committee, Statistical Physics of Cognition Workshop Imperial College London From: Rajpal, Hardik via Comp-neuro <comp-neuro(a)lists.cnsorg.org> Date: Monday, 7 October 2024 at 12:39 To: comp-neuro(a)lists.cnsorg.org <comp-neuro(a)lists.cnsorg.org> Subject: [Comp-neuro] Join Us for the Statistical Physics of Cognition Workshop – Institute of Physics, London This email from comp-neuro(a)lists.cnsorg.org originates from outside Imperial. Do not click on links and attachments unless you recognise the sender. If you trust the sender, add them to your safe senders list<https://spam.ic.ac.uk/SpamConsole/Senders.aspx> to disable email stamping for this address. Dear Colleagues, We are excited to announce that the Statistical Physics of Cognition workshop will take place on the 25th and 26th of November 2024 at the Institute of Physics, King's Cross, London. This workshop will explore the intersection of physics and neuroscience to better understand cognition, bringing together leading researchers from various disciplines, including statistical mechanics, network science, self-organized criticality, information theory, and more. The programme for the workshop is now live and can be viewed here: https://iop.eventsair.com/spc2024/programme. While registrations open next week, we invite those interested in attending to fill out an expression of interest form in advance: https://iop.eventsair.com/spc2024/registration We especially encourage PhD students and Early Career Researchers to participate by submitting abstracts for poster presentations. This is a great opportunity to present your research, receive feedback from leading experts, and engage in exciting discussions. * Abstract submission deadline: 31st October 2024 * Submit your abstract here: https://iop.eventsair.com/spc2024/abstract-submission We look forward to welcoming you to what promises to be a thought-provoking event at the frontier of cognitive science and physics. Please feel free to share this announcement with your colleagues, students, and networks. If you have any questions, don’t hesitate to reach out. Best regards, Hardik Rajpal Research Associate Organizing Committee, Statistical Physics of Cognition Workshop Imperial College London

Oct. 25, 2024

World wide VVTNS series: Fifth season opening lecture, Wednesday, October 30, 2024, at 11:00 am EDT

by David Hansel

*Important:* The seminar is at 11 am *EDT* Daylight saving time ends on Sunday, November 3, 2024 In Europe it ends on Sunday October 27, 2024 [image: VVTNS.png] https://www.wwtns.online <https://streaklinks.com/A9c7PbbpKY7PxB6PaAJWGD3-/https%3A%2F%2Fwww.wwtns.on…> - on twitter: wwtns@TheoreticalWide You are cordially invited to the fifth season opening lecture of the online VVTNS series *Karel Svoboda* Allen Institute on the topic of *Illuminating synaptic learning* The lecture will be held on zoom on October 30, 2023, at *11:00 am EDT * To receive the zoom link: https://www.wwtns.online/register-page *Abstract: *How do synapses in the middle of the brain know how to adjust their weight to advance a behavioral goal (i.e. learning)? This is referred to as the synaptic ‘credit assignment problem’. A large variety of synaptic learning rules have been proposed, mainly in the context of artificial neural networks. The most powerful learning rules (e.g. back-propagation of error) are thought to be biologically implausible, whereas the widely studied biological learning rules (Hebbian) are insufficient for goal-directed learning. I will describe ongoing work focused on understanding synaptic learning rules in the cortex in a brain-computer interface task. ᐧ -- *Il n'y a ni obligation ni pression à traiter ou répondre à ce mail en dehors des heures de travail* "В каждой шутке есть доля шутки" 'Life is good ..' (Carl van Vreeswijk, 1962-2022) --------------------------------------- David Hansel Directeur de Recherche au CNRS Co-Group leader Cerebral Dynamics Plasticity and Learning lab., CNRS 45 rue des Saints Peres 75270 Paris Cedex 06 Tel (Cell): +33 607508403 - Fax (33).1.49.27.90.62 *CONFIDENTIALITY AND PRIVACY NOTICE:* *This message and the documents that might be attached, are addressed exclusively to their(s) recipient(s) and may contain privileged or confidential information. The access to this information by people other than those designated is not authorized. If you are not the indicated recipient, you are notified that the use, disclosure and / or copying without authorization is prohibited under current legislation. If you have received this message in error, please kindly inform the sender immediately and proceed to its destruction.* ᐧ

Oct. 25, 2024

Software developer at Okinawa Institute of Science and Technology

by Erik De Schutter

The Computational Neuroscience unit (https://www.oist.jp/research/research-units/cnu) at the Okinawa Institute of Science and Technology, Japan has an opening for a postdoctoral researcher or technician to contribute to the software development of the nanoscale simulator of neuronal electrophysiology and molecular properties STEPS https://steps.sourceforge.net/STEPS/default.php. Initial duration is two years with an option to extend. The software developer will join the STEPS team and contribute to maintenance and further expansion of the software capacities. Recent versions of STEPS improved its parallel performance (http://www.frontiersin.org/articles/10.3389/fninf.2022.883742/full) and added modeling of vesicles (https://www.nature.com/articles/s42003-024-06276-5) The ideal candidate will have a scientific background but also possess good programming skills, but experienced software engineers can also apply. Competences: For the postdoctoral position a recent PhD is required in a relevant discipline such as neuroscience, molecular biology, biophysics, biochemistry, and for a technician a masters degree is required in a software/computer science related discipline. For either position, we require a talented programmer with experience in C++ and Python. In satisfying these criteria, it is expected that the candidate has some experience of modeling neuronal or other biological processes computationally, but this is not essential if the candidate has a strong enough interest in moving into the field. Fluent English conversation and writing skills are required. Benefits / Offer: A competitive salary and relocation package are offered. This an ideal position to gain experience in a prominent computational neuroscience lab. Contributions of technicians are honored by authorship on relevant publications. Please contact Prof. Erik De Schutter (erik(a)oist.jp) for inquiries.

Oct. 25, 2024

2024 Bonsai Developer Conference

by Joaquin Rapela

Hello, We are excited to invite you to the 2024 Bonsai Developer Conference, taking place from December 2nd to 4th, 2024, at the Sainsbury Wellcome Centre in London, UK. This event brings together neuroscience researchers, computational scientists, and software engineers with a shared interest in the Bonsai visual reactive programming language. The conference program and additional details can be found at https://conference.bonsai-rx.org/2024/ About Bonsai ------------ Bonsai https://bonsai-rx.org/ is an open-source reactive visual programming language widely use for neuroscience experimental control. Differently from standard programming languages where control flow is specified by composing sequences of instructions, and checking conditional loops for the occurrence of events, in reactive languages https://introtorx.com/ the control flow is specified by composing event streams. By placing asynchronous data streams front and centre, it becomes much easier to succinctly build complex programs manipulating hundreds of parallel data sources and controllers while retaining interoperability and modularity. Bonsai is a visual programming language where programs are laid out graphically. Each node represents a data stream, and edges represent input dependencies between streams. Parameters can be manipulated dynamically and every stream can be inspected at runtime. This graphical approach allows users with no programming experience to create complex asynchronous programs, and experienced Bonsai users to very quickly deploy and manage entire systems and applications. Bonsai is able to interface with a large number of experimental hardware for neuroscience experimental control, allowing researchers to easily build sophisticated experiments. We look forward to see you in London! Best regards, Goncalo Lopes g.lopes(a)neurogears.org Nicholas Guilbeault n.guilbeault(a)ucl.ac.uk Joaquin Rapela j.rapela(a)ucl.ac.uk

Oct. 22, 2024

Computational Psychology or Neuroscience Assistant Professor Position at the University of Arizona

by Cowen, Stephen Leigh - (scowen)

Hello everyone, The University of Arizona Psychology Department is hiring a tenure-track Assistant Professor in computational psychology, computational cognitive science, and/or computational neuroscience. A complete description of the position can be found below. We are looking for an individual with expertise and a publication history in domains such as modeling complex systems and behavior using approaches such as deep learning, machine learning, Bayesian methods, artificial neural networks, reinforcement learning, dynamic systems modelling, network analysis, and large language models. Preference will be given to applicants with the potential to collaborate across disciplines in Cognitive Science and Psychology. See the full posting and apply using the following link: https://arizona.csod.com/ux/ats/careersite/4/home/requisition/20668?c=arizo… The first batch of applications will be reviewed on Nov. 12. Applications will be accepted until the position is filled. Best, Stephen ------------------------------------------ Stephen Cowen, Ph.D. Associate Professor, Psychology Evelyn F. McKnight Brian Institute Interim Director: Cognition and Neural Systems Graduate interdisciplinary member: Neuroscience, Physiological Sciences, Applied Biosciences, and Cognitive Science The University of Arizona Life Sciences North, Rm 347 1501 N. Campbell Ave. Tucson, AZ 85724-5115 520-626-2615

Oct. 22, 2024

Postdoctoral Opportunity in My Research Group at University of Tennessee Knoxville

by Maroulas, Vasileios

Dear Colleagues, I wanted to bring to your attention a postdoctoral opening in my research group at the National Institute for Mathematical and Biological Synthesis (NIMBioS) and the Department of Mathematics at the University of Tennessee, Knoxville. The position is at the intersection of topological deep learning and computational neuroscience and it is a collaboration with the US Army Research lab. Please feel free to share the following link with anyone who might be interested in the position: https://www.mathjobs.org/jobs/list/25406 Thank you in advance for helping spread the word. Best regards, Vasileios Vasileios Maroulas, Ph.D. Associate Vice Chancellor Director, AI Tennessee Initiative Professor of Mathematics The University of Tennessee, Knoxville Office of Research, Innovation, & Economic Development <https://research.utk.edu/oried/> Andy Holt Tower 1331 Circle Park Drive Knoxville, TN 37996-0100 Department of Mathematics<https://math.utk.edu/> Ayres Hall 1403 Circle Dr. Knoxville, TN 37996 email: vasileios.maroulas(a)utk.edu<mailto:vasileios.maroulas@utk.edu> office: 865-974-4302 LinkedIn: Vasileios Maroulas| LinkedIn<https://www.linkedin.com/in/vasileios-maroulas-8721643/>

Oct. 22, 2024

Re: Some scientific history that I experienced relevant to the recent Nobel Prizes to Hopfield and Hinton

by Julie Goulet

Oct. 22, 2024

Re: Some scientific history that I experienced relevant to the recent Nobel Prizes to Hopfield and Hinton

by Sara A. Solla

Stephen, You have an explanation for everything. I was at the NIPS (now NeurIPS) meeting when you were giving a talk, but instead of talking about your work you spent the time allotted to you to complain about Hopfield stealing your ideas. I do not quite remember the year, but I am sure you do - late 1980s? So, yes, you have been complaining about this for as long as I know you. Word limits allow you to mention Hopfield’s 84 paper but not his 82 paper??!! It takes as many characters to write 82 as to write 84. The 82 paper was published at a time before ArXiv, preprints, and online postings, that is to say an actually printed copy became available, and that was in April 82. Your paper, the one where you did it all before, was submitted in August 82. You submitted your paper a few months after the 1982 paper by Hopfield appeared in print in PNAS. Anyone can verify this by looking at the printed versions of these two papers. What nobody can verify is that ‘someone told you that Hopfield knew about your paper’ before writing his. What cannot be verified is that there was a version of your paper that you tried to submit in 1980 to a journal that just happened to disappear. The article from 1971 that I referred to is the one you cited in your first tirade. The paper is about "Pavlovian pattern learning by nonlinear neural networks". According to Google Scholar, it has received 118 citations. And it is NOT because it appeared in a journal that has ceased to exist. It appeared in PNAS. I will assume that the 118 citations is an accurate measure of the interest that the paper has triggered in the 53 years since it was published. That you wrote a review in 1988 and that Michael Cohen wrote an article in 1992 does not give either of you precedence over Hopfield’s 1982 and 1984 papers. The Group on Statistical Physics and Nonlinear Dynamics of the American Physical Society organized a session to discuss the Nobel Prize given to Hopfield and Hinton. You can watch it on YouTube, https://www.youtube.com/watch?v=inMeFnIhuA8. My advice to you is to watch this discussion, read the 1982 and 1984 papers by Hopfield, then read your own papers, and try to see what it is that he figured out and that you missed. More than forty years later, it is time for you to look into these papers and acknowledge the crucial differences. If you still do not see it, I will be glad to try to explain it to you. Then you get to Hinton. Once I reminded you that the Prize citation is about the Boltzmann Machine and not about Back Propagation, you stopped criticizing Back Propagation to instead advertise work by James Williamson published in 1996 and 1997. These papers seem not to precede but to follow the 1985 publication of the Boltzmann Machine paper by 11 and 12 years. As for what was the atmosphere of intellectual intimidation in your group, you say that I would not know because I was not there. This is factually incorrect. I was there. I visited. I saw the group dynamics, and I was shocked. I was a young scientist, just a few years beyond my PhD, but I had never seen anything like it. One had to believe, not just agree with but believe in ART to work in your group. I ask you again: if ART provides a solution to every problem, as you advertised then and seem to continue to advertise today, how come there are still unsolved problems? The long note you have published in response to the Physics Nobel Prize is just pathetic. Sara On Mon, Oct 21, 2024 at 6:22 PM Grossberg, Stephen <steve(a)bu.edu> wrote: > Dear Sara, > > > > I reply below to your claims. > > > > *From: *Sara A. Solla <sasolla(a)gmail.com> > *Date: *Monday, October 21, 2024 at 5:12 PM > *To: *Grossberg, Stephen <steve(a)bu.edu> > *Cc: *André Fabio Kohn via Comp-neuro <comp-neuro(a)lists.cnsorg.org>, > Stephen Grossberg <steve(a)cns.bu.edu> > *Subject: *Re: [Comp-neuro] Re: Some scientific history that I > experienced relevant to the recent Nobel Prizes to Hopfield and Hinton > > You don't often get email from sasolla(a)gmail.com. Learn why this is > important <https://aka.ms/LearnAboutSenderIdentification> > > Stephen, > > You have been complaining about not getting enough credit since I first > met you in the mid 1980s. > > > > That is not true. > > > > You refer to publishing before Hopfield's 1984 paper. You deliberately > ignore his 1982 paper, which received 27946 citations: > Neural networks and physical systems with emergent collective > computational abilities, JJ Hopfield > PNAS, April 15, 1982, 79 (8) 2554-2558 > > > > My post was originally on a web site with strict word limits. That is why > I mentioned Hopfield’s 1984 article, but not his 1982 article. > > > > Section 6 of my 1988 review article: > > > > Grossberg, S. (1988) Nonlinear neural networks: Principles, mechanisms, > and architectures. *Neural Networks*, *1 *, 17-61. > > https://sites.bu.edu/steveg/files/2016/06/Gro1988NN.pdf > > > > reviews Binary, Linear, and Continuous-Nonlinear classical neural network > models by many authors. > > > > This article has been cited 2410 times. > > > > Section 6 shows that the 1982 equation that Hopfield used is a variant of > the 1943 McCulloch-Pitts model. > > > > The Liapunov function for it is just a discrete version of the Liapunov > function for the Additive Model. > > > In contrast, the 1971 student paper you mentioned received 118 citations. > > > > I assume that you mean: > > > > Grossberg, S. (1971). Embedding fields: Underlying philosophy, > mathematics, and applications to psychology, physiology, and anatomy. *Journal > of Cybernetics*, 1, 28-50. > > https://sites.bu.edu/steveg/files/2016/06/Gro1971JoC.pdf > > > > This was a review article published in a journal that ceased to exist in > 1980 (see below) and thus could not be searched well by Google, which began > in 1998. > > > > The article discussed Generalized Additive Models (see its Section 6). > > > > I was able to get articles published proving global limit and oscillation > theorems for Generalized Additive Models in the* Proceedings of the > National Academy of Sciences* between 1967 and 1971; e.g., > > > > Grossberg, S. (1971). Pavlovian pattern learning by nonlinear neural > networks. *Proceedings of the National Academy of Sciences*, 68, 828-831. > > https://sites.bu.edu/steveg/files/2016/06/Gro1971ProNatAcaSci.pdf > > > > See Equations 1, 2, 3, and 4 etc. > > > > Your 1983 paper with Michael Cohen in IEEE Transaction of Man, Systems, > and Cybernetics was received on August 1, 1982, several months after the > 1982 paper by Hopfeld had appeared in print. You did get a good number of > citations on this paper, 3344, but not as good as 27946. > > > > Michael Cohen and I tried to submit it to the *Journal of Cybernetics* in > 1980. > > > > But there was a glitch in the submission process due to the fact that the *Journal > of Cybernetics* unexpectedly stopped publishing that year and our article > got lost in the shuffle, unknown to us for a while. > > > > We resubmitted in 1982, which is why it was published in 1983. > > > > Our results pre-dated Hopfield 1982 and 1984 by two years. > > > > About the difference in citations: > > > > There is a difference between discovery and marketing. I was told that > Hopfield knew about my work before and with Michael Cohen before he went > around the country lecturing about work that we had previously published, > without citation. > > > > You mention a 1982 paper with Michael Cohen; I have not been able to find > it. > > > > I think that was a typo. Michael published the following article in 1992: > > > > Cohen, M. A. (1992). The construction of arbitrary stable dynamics in > nonlinear neural networks. > > *Neural Networks*, 5, 83 – 103. > > https://www.sciencedirect.com/science/article/abs/pii/S0893608005800085 > > > > Here is the Abstract: > > > > “In this paper, two methods for constructing systems of ordinary > differential equations realizing any fixed finite set of equilibria in any > fixed finite dimension are introduced; no spurious equilibria are possible > for either method. By using the first method, one can construct a system > with the fewest number of equilibria, given a fixed set of attractors. > Using a strict *Lyapunov function [boldface mine] *for each of these > differential equations, a large class of systems with the same set of > equilibria is constructed. A method of fitting these nonlinear systems to > trajectories is proposed. In addition, a general method which will produce > an arbitrary number of periodic orbits of shapes of arbitrary complexity is > also discussed. A more general second method is given to construct a > differential equation which converges to a fixed given finite set of > equilibria. This technique is much more general in that it allows this set > of equilibria to have any of a large class of indices which are consistent > with the Morse Inequalities. It is clear that this class is not universal, > because there is a large class of additional vector fields with convergent > dynamics which cannot be constructed by the above method. The easiest way > to see this is to enumerate the set of Morse indices which can be obtained > by the above method and compare this class with the class of Morse indices > of arbitrary differential equations with convergent dynamics. The former > set of indices are a proper subclass of the latter, therefore, the above > construction cannot be universal. In general, it is a difficult open > problem to construct a specific example of a differential equation with a > given fixed set of equilibria, permissible Morse indices, and permissible > connections between stable and unstable manifolds. A strict Lyapunov > function is given for this second case as well. This strict *Lyapunov > function [boldface mine] *as above enables construction of a large class > of examples consistent with these more complicated dynamics and indices. > The determination of all the basins of attraction in the general9 case for > these systems is also difficult and open.” > > > > I have never heard anybody referring to you as 'the father of AI'. > > > > As an acolyte of Hopfield, you wouldn’t. > > > > What I do remember is a visit to your group at BU many, many years ago - > in the late 80s or early 90s. I was amazed at finding out that you closely > supervised every word in every slide that anybody in your group was allowed > to show when invited to give a talk. Everybody was under a lot of pressure > to 'stay on message', where the 'message' was your view on things. I had > never encountered a scientific group run as a cult. It made an impression, > but not a positive one. So much for training over 100 people - by teaching > them not to think by themselves. > > > > You describe my devoted teaching as something ugly, even ad hominem. > > > > When I was a student, no one told me how to write an article or how to > prepare and give a talk or poster. I developed a method to enable my > students to grow intellectually and become independent scholars. > > > > When I started working with a new PhD student or postdoc, we discussed > what topics would potentially interest them. Then we read hundreds of > psychological and neurobiological articles together about that topic and > began to discuss the results in them. > > > > For students interested in engineering and AI, we studied all the methods > that were available to solve problems in a targeted problem domain. > > > > I encouraged my students to advance their own concepts about what might be > the underlying mechanisms that generated an article’s data or benchmarks. > Because we studied large amounts of data, favorite concepts often hit a > brick wall and had to be discarded. > > > > We developed a modeling method that I like to call the Method of Minimal > Anatomies. See Figure 2.37 in my 2021 Magnum Opus > > > > https://www.amazon.com/Conscious-Mind-Resonant-Brain-Makes/dp/0190070552 > > > > This Method is just Occam’s Razor applied to neural networks. > > > > As the Figure summarizes, it “Operationalizes the proper level of > abstraction” and “that you cannot ‘derive a brain’ in one step”. Our models > have been getting incrementally developed in a principled way to the > present time. > > > > Your statement that “the message was your [my] view on things” is the > opposite of how we worked. > > > > And how would you know? You were not there. > > > > The proof of the pudding is that all my students got positions that they > wanted after they earned their PhD’s with me. > > > > I was also told by multiple famous scientists that our department sent > them their best postdoctoral fellows. > > > > In fact, various major labs would contact me to let me know that they had > an open position and asked if one of my students was about to graduate. > > > > My students have gone on to successful careers after their postdocs and > other first post-PhD positions. > > > > Around half of them went into academe, and the other half into > engineering, technology, and AI. > > > > Groups of us have periodically gotten together for happy social events, > either at international conferences or in the Boston area. > > > > And they sent me information about their accomplishments and growing > families for many years. > > > > As for back-propagation, this is not what Hinton is cited for in the Nobel > Prize citation. It is for the Boltzmann machine. > > > > Yes, I know that, and I have also noted that my PhD student, James > Williamson, developed statistical neural network models that did not have > problems related to deep Learning. I have elsewhere noted that: > > > Restricted Boltzmann Machines (RBMS) also have analogs in the Adaptive > Resonance Theory (ART) family of models, but without the problems that > follow from using variants of Deep Learning. > > Here are two of them, both developed by my PhD student, James R. > Williamson: > > Gaussian ARTMAP: A Neural Network for Fast Incremental Learning of Noisy > Multidimensional Maps > https://www.sciencedirect.com/science/article/pii/0893608095001158 > > A Constructive, Incremental Learning Network for Mixture Modeling and > Classification > > http://techlab.bu.edu/files/resources/articles_tt/A%20constructive,%20incre… > > > > Finally, I always wondered: If ART solves all problems, why are there > ML/AI problems that remain to be solved? > > > > I would never claim that. I have always taught that science is never done. > See the Method of Minimal Anatomies. > > > > The fact that you can even think that about my work shows, sad to say, how > biased you have become. > > > > On Mon, Oct 21, 2024 at 7:49 AM Grossberg, Stephen via Comp-neuro < > comp-neuro(a)lists.cnsorg.org> wrote: > > Dear Comp-neuro colleagues, > > > > Here are some short summaries of the history of neural network > discoveries, as I experienced it, that are relevant to the recent Nobel > Prizes to Hopfield and Hinton: > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++THETHE THE NOBEL > PRIZES IN PHYSICS TO HOPFIELD AND HINTON > FOR MODELS THEY DID NOT DISCOVER: THE CASE OF HOPFIELD > > Here I summarize my concerns about the Hopfield award. > > I published articles in 1967 – 1972 in the Proceedings of the National > Academy of Sciences that introduced the Additive Model that Hopfield used > in 1984. My articles proved global theorems about the limits and > oscillations of my Generalized Additive Models. See sites.bu.edu/steveg for > these articles. > > For example: > > Grossberg, S. (1971). Pavlovian pattern learning by nonlinear neural > networks. Proceedings of the National Academy of Sciences, 68, 828-831. > https://lnkd.in/emzwx4Tw > > This article illustrates that my mathematical results were part of a > research program to develop biological neural networks that provide > principled mechanistic explanations of psychological and neurobiological > data. > > Later, Michael Cohen and I published a Liapunov function that included the > Additive Model and generalizations thereof in 1982 and 1983 before Hopfield > (1984) appeared. > > For example, > > Cohen, M.A. and Grossberg, S. (1983). Absolute stability of global pattern > formation and parallel memory storage by competitive neural networks. IEEE > Transactions on Systems, Man, and Cybernetics, SMC-13, 815-826. > https://lnkd.in/eAFAdvbu > > I was told that Hopfield knew about my work before he published his 1984 > article, without citation. > > Recall that I started my neural networks research in 1957 as a Freshman at > Dartmouth College. > > That year, I introduced the biological neural network paradigm, as well as > the short-term memory (STM), medium-term memory (MTM), and long-term memory > (LTM) laws that are used to this day, including in the Additive Model, to > explain data about how brains make minds. > > See the review in https://lnkd.in/gJZJtP_W . > > When I started in 1957, I knew no one else who was doing neural networks. > That is why my colleagues call me the Father of AI. > > I then worked hard to create a neural networks community, notably a > research center, academic department, the International Neural Network > Society, the journal Neural Networks, multiple international conferences on > neural networks, and Boston-area research centers, while training over 100 > gifted PhD students, postdocs, and faculty to do neural network research. > See the Wikipedia page. > > That is why I did not have time or strength to fight for priority of my > models. > > Recently, I was able to provide a self-contained and non-technical > overview and synthesis of some of my scientific discoveries since 1957, as > well as explanations of the work of many other scientists, in my 2021 > Magnum Opus > > Conscious Mind, Resonant Brain: How Each Brain Makes a Mind > > https://lnkd.in/eiJh4Ti > > ++++++++++++++++++++++++++++++++++++++++++++++++++++ > > THE NOBEL PRIZES IN PHYSICS TO HOPFIELD AND HINTON > FOR MODELS THEY DID NOT DISCOVER: THE CASE OF HINTON > > Here I summarize my concerns about the Hinton award. > > Many authors developed Back Propagation (BP) before Hinton; e.g., Amari > (1967), Werbos (1974), Parker (1982), all before Rumelhart, Hinton, & > Williams (1986). > > BP has serious computational weaknesses: > > It is UNTRUSTWORTHY (because it is UNEXPLAINABLE). > > It is UNRELIABLE (because it can experience CATASTROPHIC FORGETTING. > > It should thus never be used in financial or medical applications. > > BP learning is also SLOW and uses non-biological NONLOCAL WEIGHT TRANSPORT. > > See Figure, right column, top. > > In 1988, I published 17 computational problems of BP: > https://lnkd.in/erKJvXFA > > BP gradually grew out of favor because other models were better. > > Later, huge online databases and supercomputers enabled Deep Learning to > use BP to learn. > > My 1988 article contrasted BP with Adaptive Resonance Theory (ART) which I > first published in 1976: > https://lnkd.in/evkfq22G > > See Figure, right column, bottom. > > ART never had BP’s problems. > > ART is now the most advanced cognitive and neural theory that explains HOW > HUMANS LEARN TO ATTEND, RECOGNIZE, and PREDICT events in a changing world. > > ART also explains and simulates data from hundreds of psychological and > neurobiological experiments. > > In 1980, I derived ART from a THOUGHT EXPERIMENT about how ANY system can > AUTONOMOUSLY learn to correct predictive errors in a changing world: > https://lnkd.in/eGWE8kJg > > The thought experiment derives ART from a few facts of life that do not > mention mind or brain. > > ART is thus a UNIVERSAL solution of the problem of autonomous error > correction in a changing world. > > That is why ART models can be used in designs for AUTONOMOUS ADAPTIVE > INTELLIGENCE in engineering, technology, and AI. > > ART also proposes a solution of the classical MIND-BODY PROBLEM: > > HOW, WHERE in our brains, and WHY from a deep computational perspective, > we CONSCIOUSLY SEE, HEAR, FEEL, and KNOW about the world, and use our > conscious states to PLAN and ACT to realize VALUED GOALS. > > For details, see > > Conscious Mind, Resonant Brain: How Each Brain Makes a Mind > > https://lnkd.in/eiJh4Ti > > > +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > > > _______________________________________________ > Comp-neuro mailing list -- comp-neuro(a)lists.cnsorg.org > Mailing list webpage (to subscribe or view archives): > https://www.cnsorg.org/comp-neuro-mailing-list > > To contact admin/moderators, send an email to: > comp-neuro-owner(a)lists.cnsorg.org > To unsubscribe, send an email to comp-neuro-leave(a)lists.cnsorg.org > >

Oct. 22, 2024