Ricard Solé

Aquesta evolució és inevitable (no sé si desitjable) atesa la força dels interessos comercials. Encara som lluny d’una intel·ligència realment “humana”, però la IA ha permès accelerar el coneixement (no pas substituir-lo). La qüestió és si les regulacions podran limitar-ne els problemes derivats.

Why do brains generalize so well while today’s neural networks often fail outside their training data? Check this paper in @cp-neuron.bsky.social argues how neuroscience can guide better architectures & representations. Check the table below.
www.sciencedirect.com/science/arti...

Network Medicine is entering a new phase: one that demands we rethink how we study, model and ultimately treat complex diseases.

www.sciencedirect.com/science/arti...

#ComplexSystems #NetworkScience #Medicine #MedSky 🧪🧬🌐

🧵 1/

How life begins and where it might happen again A recent special issue of Philosophical Transactions B takes on one of the biggest mysteries in science: how life first began. Instead of trying to replay Earth’s exact history, the issue’s authors lo...

A new special issue of Philosophical Transactions B takes on one of science’s biggest questions: how life begins.

Rather than retracing Earth’s history, the authors look for the universal conditions that could make life possible anywhere, approaching the question from many fields and angles.

Not me.

Agreed. The rest make a lot of noise. he problem is that they are too often the ones that create public opinion about all that.

Why? Don't you realize that you are trying to provide an explanation of a large-scale phenomenon using atomic-scale concepts? There is no need for that. We have lots to do and understand but the recurrent appeal to the QM is the wrong direction (why the hell we ned that?).

I think the common attraction to physics-like views ignores the great understanding that we already have about the ways n which dynamical integration of information takes place. It's computational neuroscience, emergent dynamics and an evolutionary perspective. Reductionistic views are nonsense.

This is already happening among a broad range of disciplines, from network science, physics, imaging or biophysics to psychology, engineering and philosophy. Neuroscience IS the central part, and fairly well developped. Much to do ahead, but no magic QM is needed.

Penrose authority in Theoretical Physics has been instrumental in failing to see how weak is the "microtubule" argument. No one has shown the causal relation, and the QM effects themselves are not relevant. The Emperor is naked. Check:

I think is a really good one, with lots of estimates that sugggest that quantum decoherence dominates the small-scale neural world (and thus no causal explanation for consciousness).

I know many physicists who have moved into neuroscience (seriously, not just superficially) and are doing groundbreaking work. And I think this is awesome. And some have made strong points concerning QM and its lack of relevance to consciousness:

That's exactly what I said: "physicists WHO invoke", not ALL physicists. Tired of listening that quantum events (probably not even happening) in neurons "explain" consciousness. This is a very dishonest and arrogant view.

So after listening to multiple talks by philosophers of mind who defend consciousness as the substrate of the entire universe, and physicists who invoke quantum mechanics as the source of consciousness, the conclusion is clear: they have no idea about neuroscience, nor any interest to know about it.

Two beetles on a stick. One beetle says to the other "Do you ever get the feeling we are being observed by another lifeform" Zoom out to reveal that they are in a glass case with two scientists studying them. One scientist says to the other "Do you ever get the feeling we are being observed by another lifeform" Zoom out into space. Abouard a spaceship, two aliens watch the scientists on a screen. One alien says to the other "Do you ever get the feeling we are being observed by another lifeform" The aliens turn to look out at the reader. Text below the cartoon reads: From 'Physics for Cats' by Tom Gauld. Preorder it at www.tomgauld.

One of the cartoons in my new book 'Physics for Cats' which is out now in the UK, USA, Canada, France and Germany. Order from your local bookshops or online: www.tomgauld.com/

Big Brother is already here.

When matter came alive: the physics of life’s emergence Exploring the origins of life through the mathematical theory of transitions

Phase transitions, bifurcations, thermal vents, exoplanets and their interactive maps, as well as pointers to a whole special issue just published in @royalsocietypublishing.org about origin of life.

With Spotify and Apple podcasts as usual.

What do you need more?

CC: @ricardsole.bsky.social

I am really intrigued by your comment...

Thanks!

... which connects with the problem of individuality. Multicellular organisms (animals in particular) achieve a unique potential to develop nervous systems and learn through their lives. The multicellularity of biofilms and bacterial aggregates has a very different nature and limits.

Agreed.

I would say that the Physarum Lagrangian has the same variational structure as a free-energy functional: flows act like variational parameters, the dissipation term plays the role of expected energy, and the conservation constraints act like normalization constraints. First thoughts...

Indeed, I think these properties are too often used to make big claims about intelligence that are not supported by the observations. Bacteria in particular display remarkable collective patterns which can have adaptive meaning, but the range of behavioral responses is very limited.

There is a key point beyond the approach taken here, which is to articulate a message concerning the ongoing discussion about basal cognition and its origins. We try to make clear that the physics side of the process when dealing with a predefined graph is the right origin of the "smart" behavior.

Writing this as a Lagrangian is simply a compact variational device—it enforces “least dissipation under constraints”. The true Physarum dynamics occur in the slow evolution of the edge conductivities, while the instantaneous flow is always the solution of this constrained least-action problem.

I totally agree (and I think we made that clear) that in Physarum models, the flow at any instant is the minimizer of a dissipation functional under Kirchhoff. Our formulation does not differ from other models on networks rouring (Kelly, Phil Trans 1991) wjere a similar Lagrangian aproach is taken.

Thanls for your comments Alessandro and for pointing to your paper (great stuff). The goal here was to show that the Physarum dynamics, which has been studied using ode models onn graphs, can be mapped into a rather standard physics optmization problem onn a graph.

As you probably know, there is no universal answer (yet) for that question or for defining "cognition". We can define Intelligence is the capacity of a system to acquire information, learn from it, and use in changing environments.

Thanks for mentioning this. Is different in many ways although they share a path to solving a problem grounded in a physics minimization process.

Very good point. In fact we were considering to include in the discussion the potential extension to ant colonies and the double bridge problem, bt it turns out that, although a Lagrangian can be defined, several key things are different in a very interesting way. So... more research ahead.