On the ubiquity of power law behavior
The signature of power law behavior (short things are more probable than long ones in a predictable way) is often used as a benchmark (for instance, of a system poised at criticality).
In that pursuit, many remind us that we need to be mindful that all sorts of (not very remarkable). things can have that signature. One of the most colorful illustrations I've read is this one here:
Some Effects of Intermittent Silence
George A. Miller
1957
Imagine that a monkey hits the keys of a typewriter at random, subject only to these constraints: (1) he must hit the space bar with a probability of p() and all the other keys with a probability of p(L) = 1 - P() and (2) he must never hit the space bar twice.
Word length will fall off with power law behavior.
https://doi.org/10.2307/1419346
If you are looking to replenish this weekend, this is the thread for you. Over 30 ways to reconnect with curiosity and awe. Among them: sift through the museum backlog to find a species you've never seen before. Also: step back. Look up! And then down.
I'm bookmarking this one for a rainy day.
https://neuromatch.social/@NicoleCRust/111188025078828554
@5bafa990@2b4a17d7
(And that's why I love it here). Yes - otherwise, what's the point? Does anyone lie on their deathbed reflecting on their h-index? Hard to imagine. I certainly won't.
@2b4a17d7
Hopfield 1982 and his accompanying historical paper that explains the context: "Now what?". That doesn't exist anymore on his webpage (I've reached out to get that fixed; will email it to you).
https://www.pnas.org/doi/10.1073/pnas.79.8.2554
How do you stay connected with curiosity and awe?
I suspect most researchers experience ups and downs. I periodically find myself disconnected from curiosity and awe and I need to find that compass again. (Writing a book like I’m doing now is arguably an overindulgence there).
I worry that researchers are globally a bit disconnected from it, coming out of the pandemic coupled with increasing pressures to produce.
One impression I have is that the special swath of researchers who inhabit this furry elephant are particularly in touch with curiosity and awe.
How do y’all stay connected to it?
@fd3005de
Preaching to the choir here, no doubt, but an emphasis on dynamical systems (including integrators and attractors; bridging from intuitions to models). Not just as a pre-ANN focus but also as the present/future with a clear explanation of why (the brain is not a domino chain).
@40eabf52
Nifty!
Overall, our findings suggest that rotational dynamics and travelling waves are the same phenomena, which requires reevaluation of the previous interpretations where they were considered as separate entities.
In terms of destroys, I suspect you are referring to this?
It is important to keep in mind that the model of the rotations in the data is not a model of how the motor cortex generates volitional movements.
How brilliant is this?!
Bridging two insect flight modes in evolution, physiology and robophysics
https://www.nature.com/articles/s41586-023-06606-3.pdf
Since taking flight, insects have undergone repeated evolutionary transitions between two seemingly distinct flight modes1,2,3. Some insects neurally activate their muscles synchronously with each wingstroke. However, many insects have achieved wingbeat frequencies beyond the speed limit of typical neuromuscular systems by evolving flight muscles that are asynchronous with neural activation and activate in response to mechanical stretch2,3,4,5,6,7,8. These modes reflect the two fundamental ways of generating rhythmic movement: time-periodic forcing versus emergent oscillations from self-excitation8,9,10. How repeated evolutionary transitions have occurred and what governs the switching between these distinct modes remain unknown. Here we find that, despite widespread asynchronous actuation in insects across the phylogeny3,6, asynchrony probably evolved only once at the order level, with many reversions to the ancestral, synchronous mode. A synchronous moth species, evolved from an asynchronous ancestor, still preserves the stretch-activated muscle physiology. Numerical and robophysical analyses of a unified biophysical framework reveal that rather than a dichotomy, these two modes are two regimes of the same dynamics. Insects can transition between flight modes across a bridge in physiological parameter space. Finally, we integrate these two actuation modes into an insect-scale robot11,12,13 that enables transitions between modes and unlocks a new self-excited wingstroke strategy for engineered flight. Together, this framework accounts for repeated transitions in insect flight evolution and shows how flight modes can flip with changes in physiological parameters.
Wow! What a thread about how scientists are selected for and rewarded.
Prompted by discussion around Katalin Karikó. Quite a remarkable diversity of proposed solutions.
Warning: it's a bit heavy. But I do recommend it.
Thanks to everyone who chimed in.
https://neuromatch.social/@NicoleCRust/111167452886210356
@d4d2232e
My answer? Bend over backward to fund a more diverse range of people and ideas, even deliberately including ideas that are currently perceived as unpopular, unworkable, obscure, and the like. After all, many scientific discoveries can be traced back to origins that didn’t seem promising — like CRISPR, which began with a Spanish study on salt-loving archaebacteria in 1993 — or even to ideas that are actively opposed by the establishment.
To be sure, the success rate of this approach might be low. But if we funded 10,000 people who looked like a younger Karikó, and only one of them did something that would have the impact of her mRNA research, that would be well worth it.
How do we support the Katalin Karikó's?
There's a lot of reasonable outrage today around how Katalin Karikó was treated throughout her career (full disclosure: by my employer, UPenn). Obviously a number of someones made a huge mistake by not recognizing the brilliance and potential of her work - no question there!
What I've been thinking about and I'd love to get some scenius input on: how could we, as an academic community, do better?
Here's one summary of what happened:
https://billypenn.com/2020/12/29/university-pennsylvania-covid-vaccine-mrna-kariko-demoted-biontech-pfizer/
Taking seriously the notion that 1) we want to support the Katalin Karikó's but 2) high-risk, high-reward research takes time, here are a few ideas:
*) Better support to help geniuses communicate (and fund) their ideas.
*) More funding for high-risk, high-reward projects
*) A longer evaluation period for individuals engaged in high-risk/high-reward research
What would you add/change?
@422f8550
We're sooooo on the same page, albeit at different levels. As a brain researcher, I think about psychology (and especially psychpathology) the same way - messy!
Why complexity in biology is different from how we traditionally think about physics.
Best description I've seen:
It is often said that biological systems, such as cells, are ‘complex systems’. A popular notion of complex systems is of very large numbers of simple and identical elements interacting to produce ‘complex’ behaviours. The reality of biological systems is somewhat different. Here large numbers of functionally diverse, and frequently multifunctional, sets of elements interact selectively and nonlinearly to produce coherent rather than complex behaviours.
https://pubmed.ncbi.nlm.nih.gov/12432404/
@359dd83b
Got it! And now that I do (again, sorry to be dense): can you spell out the lesson? My sense is that it’s something like: don’t be afraid to think outside the box; even if you’re wrong, => progress!
@359dd83b@461c3d9a@2a3c13d8@6eb315a9
I love these historical examples! And sorry if I'm being dense here, but in what way was Kepler pseudoscientific? He took Copernicus's idea and Brahe's data and formulated his 3 laws; those equations made predictions about where the planets would be in the future that could be tested (And as you said, those laws triggered some of Newton's ideas about why the planets move as they do).
Is the pseudoscience here the absence of an account of "why the planets do it that way?"
In contrast, IIT appears to be prescientific insofar as most of its predictions cannot be empirically tested at all (as I understand it).
@359dd83b@14e9d346
Absolutely saddening and maddening that it's so problematic! But I'm also really thankful that we have the type of ethos in our little corner that we can discuss all this (with genuine intentions to just figure.it.out.).
Given that you found some value in the Hoel piece, I'll have a look (otherwise I wouldn't).
Elaboration of "leading theory of consciousness is pseudoscience"
from the perspective of one (of 124) author, Hakwan Lau
https://psyarxiv.com/28z3y
(For the 124 author post, see this from yesterday: https://neuromatch.social/@NicoleCRust/111074417017972359)
Among many highlights from Lau's piece:
Defines pseudoscience as
i) a set of important claims with far-reaching implications that are ii) neither currently supported by science nor are they likely to be so in the foreseeable future (perhaps even in principle), and yet they iii) masquerade as being already scientifically tested and established.
Explains why IIT is a problem as:
"Flat earth theory is clearly wrong but we don’t see articles appearing in Science11–13, Nature14, The New York Times15,16, The Economist17, NewScientist18,19, etc, repeatedly, over many years, sometimes by authoritative figures, proclaiming that it is a leading, empirically tested, and well established scientific theory. Therefore, flat earth theory is, in a sense, more ‘harmless’ and less threatening than the rise of IIT and panpsychism in the media."
Explains why IIT is pseudoscience as:
IIT shares many common features with other pseudoscientific ideas: that it is unresponsive to empirical challenges4–7; that it uses an unnecessarily complex and impractical4 language that diverges from mainstream science; that its popularity is mainly driven by the opinions by a few authoritative figures, and populist appeal8,23, rather than consensus within the scientific community or empirical success, etc. It is also notable that proponents of IIT publicly engage with religious leaders on the very topic of panpsychism24 and related metaphysical matters, and openly profess the ‘spiritual dimension’ of their ‘science’, together with controversial figures like Sadhguru25 and the New Age alternative medicine advocate Deepak Chopra26.
Also explains the events leading up to the letter and much more (for that, have a read).
Really curious to hear thoughts!
@359dd83b, @14e9d346 +++
@4ff6cf7b@064e8d4b@2f27edb2
Thanks! I’m not sure I’m getting the subtext though. There’s the idea of canalization & there’s Waddington’s valleys - both ideas/metaphors that researchers have/are finding powerful beyond Waddington’s original proposals. Wrt suitability, are you pleased or displeased with these developments?
@064e8d4b
Really appreciated! Especially given that this paper is not an easy read. The breakdown and the description of its relationship to p-factor (including the inclusion of a psyarxiv paper posted just yesterday by @2f27edb2 ++) is all really interesting and exceedingly helpful!
For anyone wanting more before the click:
The heart of the paper is the notion of “canalization.” The basic idea can be stated quite simply: it refers to how features of the mind, brain, or behavior become less able to change in a non-specific way.
Canalization serves here as a “powerful bridging construct” that connects many different ideas:
Entrenchment of a psychological phenotype
Dynamical attractors in dynamical systems
Minimization of variational “free energy” (i.e., a quantitative, statistical score of surprise or uncertainty)
...
@359dd83b
To make sure I understand (what I anticipate are very good points), your proposal is something along the lines of:
*) IIT researchers are claiming they have a theory (that's the T).
*) However, until they state what would constitute a falsification of their theory, it's pseudoscience.
Did I get it right?
@359dd83b
Great conversation! It strikes me that any declaration of pseudoscience is a prediction that an idea cannot be transformed into a testable theory. Can we ever do that in a clean way? I'm not so sure.
@359dd83b
Fair points. (I imagine some of this is about word limits). It's clearly a prototheory (not yet a theory). Whether it can ever be a testable theory is debated.
Given how high the stakes are here (the ethics of organoids, AI and coma patients (also eating salad 😉)), how wrong some journalists are getting this, and the fact that the path to this becoming a testable theory is very unclear, I endorse the term pseudoscience in this case.
That said, understanding consciousness is also one of the most challenging problems humanity has ever faced and I support the prototheory's development.
(BTW: I understand that the team will be posting responses to pushback (including this point) in a few days).
Wow! 124 brain researchers call out what journalists call the "leading theory of consciousness" (integrated information theory, IIT) as pseudoscience.
https://psyarxiv.com/zsr78/
💯: We need testable theories about the brain to move forward. Every theory starts as a proto-theory (and that's fine). But when theories are not even wrong
https://en.wikipedia.org/wiki/Not_even_wrong
we must acknowledge that.
Especially when the stakes are as high as they are here, with big ethical implications (eg for organoids and coma patients, as the authors describe).
@02ecde16
Fascinating!
Participants (N = 103) were exposed to mathematical problems without knowing that a hidden rule allowed solving them almost instantly. We found that spending at least 15 s in N1 during a resting period tripled the chance to discover the hidden rule (83% versus 30% when participants remained awake), and this effect vanished if subjects reached deeper sleep. Our findings suggest that there is a creative sweet spot within the sleep-onset period …
Stuck? Go daydream!
That's what August Kekulé did when he was trying to figure out molecular structures, including the benzene ring, in the mid-1800s. In one:
Long lines often fitted together more densely; everything in motion, twisting and turning like snakes. But look, what was that? One of the snakes had seized its own tail, and the figure whirled mockingly before my eyes. I awoke as by a stroke of lightning, and this time, too, I spent the rest of the night working out the consequences of the hypothesis.
And that's how the idea of the benzene ring was born.
https://www.nature.com/articles/465036
@c4c92358
Great question! Analogous to Box & Draper’s “all models are wrong but some are useful”, that applies to descriptions too. When does the wrongness of a description cross the line into a myth that’s worth busting?
If individuals want to dopamine dress or dopamine detox, I applaud the spirit (even if I don’t agree on the mechanism). Who am I to rain on that party?
My take is that it is my role to 1) spread sense not nonsense (silence across the board is suboptimal; fill the world with good content! 2) do it in a way that is not obnoxious.
@8a2094ca
Great point!
By absolutely sheer coincidence (and definitely not because I'm any type of well read fancy pants), I was just diving into another old book whose last chapter you might be interested in:
The wisdom of the body.
Walter Cannon, 1932
(This book was the first to coin the term "homeostasis" and introduce fight-or-flight).
Available via the internet archive.
https://neuromatch.social/system/media_attachments/files/111/024/397/564/858/447/original/417b455f5e2534fb.png
A wow prescient quote from 1865 that will resonate with my anti-reductionist friends
“Physiologist and physicians must never forget that a living being is an organism with its own individuality. Since physicists and chemists cannot take their stand outside the universe, they study bodies and phenomena in themselves and separately, without necessarily having to connect them with nature as whole. But physiologists, finding themselves, on the contrary, outside the animal organism which they see as a whole, must take account of the harmony of the whole, even while trying to get inside, so as to understand the mechanism of its every part. The result is that physicists and chemists can reject all idea of the final causes for the facts that they observe; while physiologists are inclined to acknowledge a harmonious and pre-established unity in an organized body, all of whose partial actions are interdependent and mutually generative. We really must learn, then, that if we break up a living organism by isolating its different parts, it is only for the sake of ease in experimental analysis, and by no means in order to conceive them separately. Indeed, when we wish to ascribe to a physiological quality its value and true significance, we must always refer to this whole, and draw conclusions only to its effects in the whole.”
Stumbled upon via this excellent take about homestasis (and how it's anti-reductionistic)
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7076167/pdf/fphys-11-00200.pdf
Claude Bernard, 1865
@e298e1ac
Absolutely! And thanks for the article.
The parallel pursuit of treatments at higher levels as we chase cures (and treatments) at lower levels should happen, for sure. I recently had a very insightful conversation with an individual who has a rare degenerative disorder. A component of her patient advocacy is spreading the word to researchers that the pursuit of cures should not happen at the expense of pursuing treatments because treatments can vastly improve quality of life (and cures seem to be decades away).
In my mind, there is no singular right approach here. The notion that we (as a community) pursue multiple approaches in parallel (what some like to call "pluralism" and others like to call a "distributed portfolio") is the only thing that makes any sense. We need to play both the short and long game at the same time.
Of course each of us, as individual scientists, needs to make the decision about what we want to pursue ourselves (what we are most optimistic about and interested in, etc).
In my mind, tackling the degeneracy is probably the only route to cures. Whether this happens in my lifetime or not is very much TBD. What we do know is that we definitely won't make any progress if no one works on it.
@e298e1ac
The emerging story for nearly all brain disorders is that the same symptoms follow from a plethora of causes and different individuals with the ‘same’ disorder will require different treatments (precision medicine). I believe that the degeneracy you are talking about here is closely related.
Cystic fibrosis (while not a brain disorder) is one example - it’s always a mutation to one gene, but the drug needed to treat it depends on what the mutation is. Now take that complexity and multiply it for the more typical disorders that aren’t associated with a single gene (but many genes + environmental interactions).
So yes on appreciating the degeneracy and understanding those many biophysical stuffs to one function principles!
I dipped a toe into Bluesky for a bit and I'm not impressed. A lot of quirky trivia but not so much interaction happening there. To each their own! I have a small pile of invites; if you want to see what it's like, PM me and I'll be happy to pass them along.
@e298e1ac
CNNs are great for some things. Epilepsy, Alzheimer’s and depression are probably not on that list. Biophysical models, however, likely will be impactful for those problems.
Oddbean new post about login | logout
Notes by Nicole Rust | export