iiinova — where your digital self keeps evolving

Toviah Moldwin @tmoldwin.bsky.social · 17/05/2025

In the grand scheme of things the main thing that matters is advances in microscopy and imaging methods. Almost all results in neuroscience are tentative because we can't see everything that's happening at the same time.

110

I also have a stack of these, I call it 'apocalypse food'.

But so is every possible mapping,so the choice of a specific mapping is not contained within the data. Even the fact that the training data comes in X,y pairs is not sufficient to provide a mapping that generalizes in a specific way. The brain chooses a specific algorithm that generalizes well.

(Consider that one can create an arbitrary mapping between a set of images and a set of two labels, thus the choice of a specific mapping is a reduction of entropy and thus constitutes information.)

100

The set of weights that correctly classifies images as cats or dogs contains information that is not contained either in the set of training images or in the set of labels.

200

Learning can generate information about the *mapping* between the object and the category. It doesn't generate information about the object (by itself) or the category (by itself) but the mapping is not subject to the data processing inequality for the data or the category individually.

Toviah Moldwin @tmoldwin.bsky.social · 15/05/2025

GPT is already pretty good at this. Maybe not perfect, but possibly as good as the median academic.

100

What do you mean by 'generate information'? What is an example of someone making this sort of claim?

Toviah Moldwin @tmoldwin.bsky.social · 13/05/2025

Paying is best. Reviews should mostly be done by advanced grad students/postdocs who could use the cash.

Toviah Moldwin @tmoldwin.bsky.social · 07/05/2025

Why wouldn't you want your papers to be LLM-readable?

100

Toviah Moldwin @tmoldwin.bsky.social · 04/05/2025

If such a value to society exists, it should not be difficult for the PhD student to figure out how to articulate it themselves. A lack of independence of thought when it comes to this sort of thing would be much more concerning.

Toviah Moldwin @tmoldwin.bsky.social · 04/05/2025

Oh you were on that? Small world.

But I do think in our efforts to engage with the previous work on this, we made this paper overly long and technical. We present the bottom-line formulation of the plasticity rule in the Calcitron paper.

A generalized mathematical framework for the calcium control hypothesis describes weight-dependent synaptic plasticity - Journal of Computational Neuroscience

One of the reasons we wrote this paper is because the calcium control is a great theory but there were two semi-conflicting mathematical forumlations of it, both of which had some inelegenancies. I think we managed to clean them up, and made it more 'theory'-like. link.springer.com/article/10.1...

link.springer.com

The brain modifies synaptic strengths to store new information via long-term potentiation (LTP) and long-term depression (LTD). Evidence has mounted that long-term synaptic plasticity is controlled vi...

I know that e.g. Yuri Rodrigues has a paper that incorporates second messengers but at that point it's not really parsimonous any moren

110

The leading theory for plasticity is calcium control, which I've done some work on. I do think that I've contributed on that front with the Calcitron and the FPLR framework which came out in the past few months. Anything beyond calcium control gets into simulation territory.

220

The reason why it's less active now is because people kind of feel that single neuron theory has been solved. Like the LIF/Cable theory models are still pretty much accepted. Any additional work would almost necessarily add complexity and that complexity is mostly not needed for 'theory' questions.

Hebbian learning? Associative attractor networks (e.g. Hopfield)? Calcium control hypothesis? Predictive coding? Efficient coding? There are textbooks about neuro theory.

050

I kind of like the size of the single neuron theory community, it's the right size. The network theory community is IMHO way too big, there are like thousands of papers about Hopfield networks, that's probably too much.

140

Toviah Moldwin @tmoldwin.bsky.social · 02/05/2025

Not really true, there are a bunch of people doing work on e.g. single neuron biophysics, plasticity models, etc. Definitely not as big of a field but we exist.

150

And 'because it's unethical in this situation' is not a valid response; the ethics are irrelevant to rigor and the epistemic question of the scientific approach to establishing truth in medicine.

000

Toviah Moldwin @tmoldwin.bsky.social · 01/05/2025

Exactly. There's no difference between a 'peer reviewer' appointed by a journal and you, a scientific peer, evaluating the publication yourself.

Toviah Moldwin @tmoldwin.bsky.social · 28/04/2025

Ariel Krakowski interviews me on his podcast about brains, AI, plasticity, connectomics, consciousness, and everything in between. open.spotify.com/episode/4m33...

open.spotify.com

Computational Neuroscience, Connectomics, and Consciousness

Zappable · Episode

Toviah Moldwin @tmoldwin.bsky.social · 27/04/2025

If they're actually that great there's no problem, the university will want to keep them. But if their job would be at at risk from a younger competitor if not for tenure, that's evidence that they're not actually that great.

000

Toviah Moldwin @tmoldwin.bsky.social · 27/04/2025

Also a lot of professors build their research careers on outdated scientific trends, in science especially you don't want something holding you to the past.

000

Toviah Moldwin @tmoldwin.bsky.social · 27/04/2025

Most professors, including tenured ones, are not 'great minds'. But either way, there's no reason to think that the average tenured professor is a 'greater mind' than the best applicant from the pool of recent grads who could potentially replace them.

200

I agree, it's even more important for public goods like science to not have incentive structures, like tenure, that are counterproductive to the public interest.

000

Tenure is a large part of the reason science advances 'one funeral at a time' when it should be advancing one hiring/firing decision at a time.

000

Jean-François Cudennec @limpetnerd.eurosky.social · 26/04/2025

No. Again, having tenure disincentivizes productivity for those who have tenure and makes things harder for younger academics. It also means old ideas and people become entrenched instead of letting in fresh new perspectives to compete with them.

100

Reposted by Toviah Moldwin

On today’s science headlines : A newly discovered species of carnivorous caterpillar on the Hawaiian island of Oahu has earned the nickname “bone collector” for its eerie habit of adorning itself with the remains of its prey, such as ant heads and fly wings. 🪲 🧪

Headline of Rubinoff et al paper : « Hawaiian caterpillar patrols spiderwebs camouflaged in insect prey’s body parts »

Editor’s summary of the Rubinoff et al paper

Examples of prey’s body parts used as camouflage

171

To which situation are you referring? The binary adjunct/tenure distinction? According to GPT it started when American universities started to expand in the 1940s, especially because of the GI bill. Before then basically everyone got tenure by default but it was less formal.

100

From the perspective of a PhD grad/postdoc looking for a job-would you rather the job market be A) a variety of positions at different pay for different levels of experience, with possibility to advance or B) an all-nothing system where a tiny % of people get a tenure job and scraps for the rest?

100

People would like job security in other industries too, but they don't get it because permanent job security generally doesn't makes sense. Yes it's a perk but it's a market-distorting perk.

000

The category of 'adjunct' only exists because tenure exists. Abolish tenure and you'll just have normal positions on a gradual payscale like every normal industry.

300

It also creates a two-tiered system for hiring instead of a graded one, which results in all sorts of weird distortions to the academic job market.

100

Tenure broadly is probably a bad thing. Disincentives productivity, entrenches older people in positions instead of allowing younger people to compete with them, etc. It doesn't exist in any other industry and someone would have to make a case for it in academia beyond historical reasons.

220

But what's the alternative? This is basically how all molecular biology works, it's a network of molecules that regulate other molecules via various intermediaries. Is there some other experimental paradigm?

000

Sure, but, like, you can do experiments that demonstrate mediation. I.e. perturb A and observe that it changes C, then you block B and observe that perturbing A no longer changes C.

110

These terms are used in biology a lot because this is just how biology works. It's a big network of molecules that all affect each other in dose-dependent manners. This dose goes up, more of this thing is made, which stops that thing from being made. That's how it is, we didn't make the rules.

050

People may use these terms incorrectly but they have well-defined meanings in causal inference and should be used when appropriate.

150

Toviah Moldwin @tmoldwin.bsky.social · 22/04/2025

Most of these are fine in many contexts. Modulate means to perturb, i.e. 'do' in the do calculus. X regulates y means perturbing X changes Y. Mediate is well-defined in causal inference also, i.e. if B mediates between X and Y, Perturbing X changes B which changes Y. Etc.

160

Join us in Jerusalem for a great program!