Deciphering the patterns of nature is something that has occupied curious minds for countless generations, from swarms of bugs to the flight of birds to the movement of your bodies own cells.
Prof Andrea Cavagna of the Institute for Complex Systems leads a research group bringing together the worlds theoretical physics and experimental biology, building mathematical models of natural systems and uncovering the universal laws that underly the organisation structure of life.
Read the original paper on swarm behaviours: https://doi.org/10.1103/PhysRevLett.123.268001
This is an automatically generated transcript.
Hi, I’m Will. Welcome to researchpod.
Deciphering the patterns of nature is something that has occupied curious minds for countless generations of better understanding of the mathematical laws. Underlying those patterns is something that has only recently become available and have seen accelerated growth with new tools and technology.
Professor Andrea Cavagna of the Institute for Complex Systems leads a research group bringing together the worlds of theoretical physics and experimental biology, building mathematical models of natural systems, from bugs to birds to your body’s own cells, and uncovering the universal laws that underlie the organisational structures of life.
For my own sake and for everyone listening at home, could we start with hearing about what it is you’re currently working on?
00:01:03 Prof Cavagna
I’m currently research director at the Institute for Complex Systems, which is part of the Italian National Research Council here in Rome.
I’m a physicist. I have a PhD in practical physics and I spent the first 10 years of my career between Rome, Oxford and Manchester and then back to Rome again working on problems in statistical physics and condensed matter.
And could you tell us a little bit more about what that is? Hopefully using some familiar terms from everyday life.
00:01:37 Prof Cavagna
Broadly speaking, statistical physics studies system made of large number of particles, large number of units which interact with each other through some relatively simple physical laws.
They give rise to macroscopic emergent behaviour that cannot simply be guessed by the microscopic details of the particles.
One of these emergent behaviour is collective behaviour, so there are cases in which very many unit interacting with each other give rise to macroscopic patterns on large scales, and this will what we call collective behaviour. And I’ve been interested in collective behaviour basically for my life, even though in the first part of my career I focused on the collective behaviour or physical system.
And this is what convinced Martin studies rather than the collective behaviour of biological system, which is actually what I do right now.
When it comes to bringing together physics and biology like this, or do you say is the interest or the motivation that you have in the topic?
00:02:42 Prof Cavagna
But you know, we turning from theoretical physics which what I was doing before 2 experimental biology was a very great change of course, and was a step that was also somewhat problematic for my career.
On the other hand, there was natural something natural in this change, because my research.
Which had been very much theoretical for many years, and at some point I probably felt that the direction I was going along with perhaps too abstract, kind of excessively detached from anything I could actually see or experience in my life.
But I really like the concept involved in collective behaviour and emergent phenomena in general, so I was probably longing for a change, but I change that did not bring me conceptually too far from this topic.
And then there were flocks so Starling flocks are really. Amazing abundant here in Roma. Extremely imposing during the whole winter months.
And their evolution patterns are truly beautiful.
Me and everybody in Rome is spending the best part of the sunset minutes by looking at them with your neck up to the Sky. So together with some other colleagues I was working with at the time, especially Georgia, Breezy and they’re injured.
We thought that perhaps we could use some of the knowledge or the tools that we had developed for the study of collective behaviour in physical system.
But to study biological assistance, in particular, this wonderful flocks that we had in front of our very eyes everyday at sunset. And this is what we did in 2005. We asked for some grant money to the European Commission to start the project for the experimental and theoretical study of the collective behaviour in Starling flocks.
European Commission graciously granted us the fans, and in fact it has been funding our research on this topic ever since, so the first project gave us the resources to do this, and the project was quite successful. And in the course of the last 15 years, it expanded its boundaries quite a lot. So now we started a collective behaviour not only in Starling flocks, but also in insect swarms, and we are starting new experiments. Also on send colonies and others systems. But all of this is done with an approach which heavily uses the tools and the concept of statistical physics.
So in this sense, after all, I haven’t changed much my my topic of research.
Which leads us neatly to the RG BIO group today.
00:05:35 Prof Cavagna
Yes. So when when we started this study to 15 years ago, What we had in mind was mostly to do models and theories of collective behaviour in by a system.
But we immediately realised two things.
The first thing was kind of disappointing. We realised that biologists, physicists and computer scientists.
They had already done a lot of work on modern collective behaviour, especially bird flocks and mammal heard, so we were definitely not the first one having this idea. For example, Craig Reynolds was a computer scientist doing computer animations or flock as early as 1986.
For 1985 I think, and some of these animation were actually very beautiful and they were even using the in the movie industry so.
Okay, we may have been smart, but certainly we were quite late in joining the field of modelling of computer model of collective behaviour. But the second thing we discovered was more encouraging.
Also, a bit weird because we discovered that there was a great variety of models and theories, but very, very little experimental data.
Especially in three dimension, where flocking actually course and that was you, of course. Today’s significant technical challenges of taking 3 dimensional data of large flocks of birds in the in the in the wild.
So there had been some pioneering studies in I think it’s late is back in the 70s, especially by Frank Heppner.
But they involved very small flocks, up to 810 pigeons. While a statistical physicists who were interested in large groups, there is a venerated article by Phil Anderson, was a Nobel Prize in condensed Matter Physics, which is famously titled more is different.
More is different, meaning that that emergent phenomena really require large number of particles or units or animals to be qualitatively different from their individual counterparts.
So what we decided then was to get out in the field and collected. Swear mental data on large flocks of Starling.
Choose what we did so in this way I turned from theoretical physics to experimental biology, sort of overnight.
And collecting this data on flocks of Starling took us very long time, even though we could use digital photography compared to people in the 70s, we still had some very big challenges ahead of us. The method we use is basically stereoscopy which is more or less what our eyes do to estimate the three dimensional position of objects in front of us. We take synchronised images of our targets.
00:08:30 Prof Cavagna
In this case, the Birds within the flock from different point of view, so from different cameras we have a set of three Cameras and then we use these images to reconstruct the full 3 dimensional position of each bird. There velocities and whatever else we want to measure.
Well, this is easier said than done, and it requires us to develop some novel non trivial computer vision techniques. But it was worth it because.
The experimental data or what really made the difference in the first years of our studies and experiment was really what put us on the map in the field of collective behaviour. However, along the years we also started doing some new theories, building on our experimental day.
And so our theoretical approach expanded, and eventually two years ago I was awarded an advanced grant by the European Research Council.
For a project called RG Bio, in the context of which we apply while we try to apply one of the most fascinating tools of statistical physics, which is called the renormalization group to the study of collective biological systems.
So part physics, biology, part mathematics and part wildlife photography all coming together to make this free normalisation theory like where would this fit in in what other people might experience or understanding of research and academia?
00:10:11 Prof Cavagna
To understand what the Renormalisation Group is , or RG, and so to understand what the RG is and how it can be used to uncover some of the important phenomenon collective.
Here one has first illustrate two crucial aspects to creature concept which are common to statistical thesis also which are Interaction and correlation.
So interaction correlation are really the two pillars of stat fees and condensed matter, and I also think that they are the Two basic pillars of collective behaviour in biological system.
So once we had the date that the experimental data, the first question we asked was how do Starling’s interact with each other within a flock? So clearly they do interact because their motion. This guy is wonderfully coordinated and very far from random. So how do they do that?
So by interaction here, I really mean what kind of rules do they use when keeping each other under control to get around so well together.
To previous models and theories had hypothesise that in fact there were some very simple rules at work. Simple rules that could be enough to grant their collective code initial absurd and the rules are most intuitive. Namely align your direction of motion to death of your neighbours so you don’t crash into the other guys. Keep a distance, but not too much of a distance, otherwise you get too far from the group, so these are very, very simple rules. Alignment and keeping a distance which is not too small and not too large.
00:12:03 Prof Cavagna
So our data showed that these rules were basically sound very sound, but we also discover something quite new, Something that now we call topological interaction as opposed to metric interaction.
The fact is that all kind of interactions in nature and special in physics they decay with the distance particles or planets or whatever that are at a very large distance, interact more weekly, and if they are closer.
So this decay also happened to course in within a flock of Starling’s. The interaction is stronger the closer the neighbours are do each other.
But the crucial point is that while in physics this decay of the interaction depends on the actual physical distance metric distance between the particles, so whether they are one metre apart, 10 metres apart 100 metres apart.
What we discovered in Flock is that it is not like that If you imagine that you are one bird in the flock, you’re not interacting with all neighbours within given fixed distance.
But what you do is that you’re interacting with a fixed number of Neighbours.
00:13:25 Prof Cavagna
And this number is between 7 and 10 on average. So this is something very peculiar which does not happen in physics.
So it’s like birds were measuring the distance between individuals, not in unit of metres, but in unit of other individuals And this is what we call topological distance . And so a topological interaction is based on this kind of distance.
Even though this is certainly not something that you find in a in a physical theory because it requires some neural activity by part of the the animal which is doing that because it’s keeping under control a fixed number of neighbours is something quite natural. When we think about the biological or even the social assistance.
For example, when we are in the subway, when we use the underground, we do not measure the distance we have to travel in unit of metres, but in unit of stop. So we count how many stops there are between this and this other place, not how many kilometres there are or if you consider a network there relevant distances, the number of nodes between two given notes.
So this concept of topological distance is quite natural when you go outside physics, and so it was funny that actually previous even biological models trying to describe this logical system were based.
On a strictly physical concept of interaction which was decaying with physical distance. While what our data showed is that it is not like that.
Sounds less like physics and more social networking.
00:15:10 Prof Cavagna
Yes, exactly this seven neighbours with which they are interacting. I really the one nearest neighbour. The closest to the focal bird, so it’s not that they’re interacting with seven guys that could be wherever they want into the flock. Now they’re interacting with a fixed number of neighbours, which are the closest neighbour the nearest neighbours.
But their actual distance doesn’t matter. So if you have a flock which is expanding, mutual relative distances are increasing.
The strength of interaction remains exactly the same because it doesn’t depend on the actual distance between me and my neighbours because I’m only keeping under control. If it’s a number of guys.
First shell of neighbours around myself irrespective of our distance they are.
So in this sense, this interaction is not metric. Is topological, however, still very much local in space?
I know flocks can be huge formed up 200,000 birds, so it’s not that in a flock of 100,000 Starling, so they’re interacting everybody with everybody else, no, not at all.
Every bird is interacting only with a very relatively small number of neighbours, against something between 7 and 10.
So then our next question was how is this possible to achieve such long range coordination? Such huge and macroscopic coordination through such a short scale local interaction?
So this is where correlation comes into play, which is the 2nd crucial concept in our path.
So while interaction is a direct transfer of information between two or more individuals.
Correlation is an indirect transfer of information.
So two birds may be correlated to each other and therefore their behaviour may be similar to each other even though they have never interacted with each other.
A simply because there is a chain of interaction between those two birds which make them behave similarly even though they have never interacted with each.
And this is something which of course happens all the time. Also in social systems and as well as in physical system. So correlation is due to the indirect transfer of information through this chain of interaction.
A nice example of that is the phone game that children play, so each child whispers a message in the year of the nearby child, and that is a direct interaction. So each child is just interacting with one other child, and then the message travels travels through the whole chain of children. At some point it gets corrupted because there is noise in this channel of communication.
But the distance by which the messages travelled can be larger than the one child interaction range of community of direct communication. So this larger span of the travelling with information is what we call the correlation length. So the extra space extension of the correlation. Correlation is due to interaction is connected to interaction, but is something different.
So you can have very short range interaction, but a bit larger range correlation,
What we did with our data was to actually measure the correlation function in Real Flock of starlings. So we really measure how much they behaviour the direction of motion of 1 bird at some point is influenced by the change in the direction of motion of another birds 20 metres away.
We completed that and what we found was something quite surprising, so we found that this range of the correlation was as large as entire flock.
So this is what we call in physics scale free correlation, because mean that there is no actual correlation length and they only scale of the correlation is the size of the system itself, so it gets as large as it can. It cannot be larger than that because there is nothing larger than the system itself within the flock.
No upper limits or boundaries, just as far as the reach extents.
00:19:44 Prof Cavagna
Exactly, so this is something which can happen also in physical system, but it’s something quite rare.
You have to normally tune many parameters in your system in order to have a correlation length, which is as large as the entire system.
So that were the two great pieces of experimental phenomenology we got from our data, very short range interaction but very long range correlation.
On one hand, these kind of explained how they can get this fantastic coordination of the microscopic scale. Anyway, it’s a it’s a technical definition of what do we mean by saying that the whole is more than the sum of its parts, which is something.
00:20:37 Prof Cavagna
The nice catchphrase that you listen a lot when you study collective behaviour, but it’s never quite clear what it means.
Well, given what we discover, it really means that because the correlation is as large as entire system, so parts of the system which are very far from each other are still correlated to each other. If you now divide the system in smaller sub part, if you really take them apart from each other, there are different because you’re killing that correlation.
So in that respect, is the long range scale for correlation, which makes the system the whole system more than the mere some of its part.
On the other hand, this is cover the short range interaction accompained by long range. Correlation was really what justified us to make the next step, which is what this current project of ours is about. It is to go something towards something more ambitious. Is the scaling hypothesis there re normalisation group and eventually to universal.
And when you say universality is that a universal set of laws are universal modelling of behavior.
00:21:58 Prof Cavagna
By universality, mean that you can have cases in which different microscopic set of rules can end up causing the same macroscopic behaviour of the system.
And this is something very profound that was discovered by the physics of the last century, which is really what we would like to export to biological system. The scenario described before, in which short range interactions give rise to long range correlation they prefer.
At that, we discovered that in flock of stylings was also great paradigm within statistical physics. In the last sent.
It was really the pillar of condensed matter physics.
And some of the great scientist that time were prompted by this scenario to formulate very very bold hypothesis very bowl for those years, which is normally called descaling ipadis. So this is part of this states that whenever the conditioning systems are such that long range correlations emerge from short range interaction then Microscopic details, Teeny tiny details do not matter anymore.
And this was something truly revolutionary at the time, because after all, microscopic details are what actually define an differentiate the system with respect to one another. So how can they not matter anymore when correlations are so long range? And so when the correlation spanned inter sets?
It’s like the system emancipate from its microscopic constituents and its behaviour then only depends on a handful of very general things like the dimensional space, the symmetries of the system, the conservation loads, but not the actual small details defining the dynamics.
The important point is this. Killing hypothesis is not only something very beautiful from a theoretical point of view.
But it’s also incredibly useful and practical from the experimental point of view, because If the microscopic details don’t matter anymore.
00:24:15 Prof Cavagna
Then systems apparently different from each other. They actually end up having the same macroscopic, the same large scale behaviour.
And this is what really happens in physical systems and is what is called universality.
And the RG their realisation group Was the way in which this hypothesis was actually proved was mathematical and physical proved to be correct?
Well, with reference to the telephone game that you mentioned, is there a chance for erosion or decay in the message that the signal going in might become corrupted by the time it reaches the stage of manifesting in behaviour? I’m thinking of what people listening at home might think of as chaos theory that even under ideally reproduced circumstances, there might be some subtle or undetected shift or change in circumstances that leads to a different end, but.
00:25:12 Prof Cavagna
I. wouldn’t say that this is that the chaos paradigm is relevant in in this particular system, actually quite the opposite. I would say that they one of the great features of collective theological system is their great robustness.
So system there is noise. There are allsorts of noise that is noise due to the external environment. There is noise because no biological being can be able to perfectly implement some set of behaviour rules. So there all sort of noise.
However, the overall microscopic behaviour of the system is very stable against this noise, so I would say that probably evolution worked towards making this kind of system very stable against chaos.
So even though there are definitely, there is definitely room for no linearity is and non linear physics in these systems. I would not say that chaos is the right paradigma.
So how do we get from these Sometimes microscopic details out to a universal law?
00:26:27 Prof Cavagna
The central idea of the RG is that, of course, graining integrating out erasing. Anyway, information about a system at the short scales.
And see how this integration modifies the behaviour of the system at the largest case.
It is this operation of changing scales iteratively, going from shorter to increasingly larger landscaped, increasing the largest distances.
That is regulated by the by their normalisation group. So the idea is that at the beginning when you are still at the level of the small microscopic details, really they short scale interaction rules.
Then they re scaling procedure, changes the behaviour of the system and a lot. Once this course training reaches the macroscopic level, the behaviour does not change anymore. So in this sense they system is at what we call a fixed point of the randomization group.
00:27:36 Prof Cavagna
And the crucial thing is that through this process, the physical system you’re studying, which is a sort of basing of attraction To which other system belongs so that season that had very different microscopic details that were starting in this?
Crosses from very different, less a position. The parameter space. All this system in fact flow towards the same basing of attraction, so they float towards the same physical behaviour through their normalisation group.
The fact that many different microscopic system, which were apparently very different from each other.
Flow to the same microscopic physics to the same microscopic behaviour is what we call universality, which was hugely powerful concept. This was a tool.
Thanks to each people understood why in the in the 22nd why system as different as the liquid vapour transition or actually described by exactly the same physical laws that was describing Ferromagnets.
00:28:50 Prof Cavagna
There was no connection whatsoever at the microscopic system between, you know the the liquid vapour transition. So the world of condensed matter and liquid, and so one. And the physics of ferromagnets, but people were obtaining experimentally the same answers over Andover again.
And finally, the RG explained what it was going on because those systems when you start integrating out short scale details By re scaling the system that is watching the system at larger and larger scale.
00:29:29 Prof Cavagna
You end up in the same basing of attraction with the same exact physical behaviour at the largest case. The RG is also not only constable is also a practical tool by which you can compute physical quantities characterising assistant.
And it is clear then why it could be very useful. So in biological system We found this phenomenology of having large scale correlations. Not all in flux.
I don’t see an apparently very different system like swarms, swarms of insects in that case, to you have a short range interaction but long range correlation and other people. Other groups I’ve found a similar behaviour is similar phenomenology. In other system bacterial clusters. And many different systems in biology.
So if one could apply the same tools of universality and their normalisation group in biology, that would be extremely useful, because the diversity of physics is really nothing compared to the incredible diversity of biology, so the role of small details it’s really are.
Really arresting it biology, but if something like the rumour is Action Group.
Can be used also in biology. Then perhaps we can just hand a little bit all these microscopic different details and just classify biological systems into perhaps few universality classes with the same behaviour.
For more about this research in action. I spoke with Dr Giulia Pisegna from the argie bio group about a paper from the physical review letters looking at the renormalization group approach to swarm behaviour in insects and they started off by asking Julia what it was that she got from the work in the research that she does.
00:31:28 Dr Giulia Pisegna
Basically, what excites me to do our line of research is that of course. You you start with a problem and the problem here is an evidence that comes from experimental data.
And then you try to make the best, but also the simplest choice from a theoretical point of view to see if the theory can relate to the experimental data.
But anyway, at the beginning is just a tentative, so you have some kind of intuitions, but you don’t know if that intuition will be correct.
So the result that now we explained we are explaining in this paper starting from the fact that a Andrea and also the other seniors tried to apply that specific model to the problem as worms, but they were not really convinced that that model would have worked to really give a value consistent with experimental one.
At the beginning I did some numerical simulations of this model, so some investigation about that model and I found a value that was more or less a surprise for me, but also for all the other people that were participating to the research.
And it was really exciting because that result was confirming that the intuition at the beginning was.
Wrecked and hold so that we were on the right path to really hear. I’ve tried radical explanation of living phenomena that seems very far from Henny mathematical and physical description, so I was feeling a very good sensation to say, okay, I’m doing something correct and something that really can explain something that was so unclear to everyone and so.
Full of mystery, because it’s a phenomenon of life, basically.
00:33:45 Dr Giulia Pisegna
One of the most exciting things of this field is that now physicists are trying to transfer their knowledge and methodology to the study of living and biological systems.
On the one hand, we are all used to think about the physics of inanimate matter has, for instance elementary particles or new states of matter.
On the other, it is a bit unusual to think about the physics of insects, worms, or flocks of birds. However, their research of our group.
Revealed that this natural living system, that one can see everyday flying in the Sky or swarming in the parks, and that seems also very far from any kind of physical and mathematical description.
In fact can be studied using the same mathematical and physical tools of classical statistical physics.
An what like of this kind of research is the general approach of the new and modern biophysics, which is trying to say that as there exists, come on mathematical laws to describe phenomena, has the motion of planets or electromagnetic waves. There could exist also an underlining mathematical structure.
Yeah, in living natural phenomena that has yet to be discovered Insect swarms are one of the biological system on which we focus our research.
Our group had been able to collect many experimental data in the past years for these systems that is warm in their natural environment of the packs of room.
By using the same technology to record the trajectory’s of each individual in the group already applied to flexibles.
What we discovered, starting from this data is that the motion of the system is not purely random or DIS organised. Indeed, insect swarms show the same fundamental ingredients of collective behaviour that we found also in flux, namely short range interaction of alignment of their direction of motion. And also are strong long range correlation that spans all the linear size of these groups.
This basically means that even for neutral swarms, the behaviour of an element of the group can be influenced by the behaviour of another element that can be hold so very far from it.
And with this system is quite different from the one a flexible.
00:37:01 Dr Giulia Pisegna
First, because we are Speaking of different classes of animals, but also because their phenomenology is quite different.
In the sense that swarms don’t show a global coordinated motion as birds do.
But a weather disorder one around a specific point of this space.
Anyway, despite this difference, the two systems seem to display the same characteristics that define collective behaviour.
And we think that this fact could really suggest the existence of a possible common theory able to describe both of them.
So how are recent research started from the experimental evidence?
That natural swarms follow our fundamental law of classical physics systems that we call the dynamic scaling. So this property says that if we look at devolution in time of are strongly correlated system.
Microscopic details of the system itself don’t count allot. Also, in determining the dynamical macroscopic behaviour of this system.
The fact that these systems display collective behaviour, it means in a sense that we can forget about the tiny details of these animals and of the particles that compose this system. For instance, if we are looking at images or.
Or if you are looking at birds, we can neglect the fact that they are particular shape or that they fly in a particular way.
We can neglect hold this kind of also let me say biological but also important very important details to simplify as much as we can Howard description and so try to use the same model for systems that are apparently very far from each other. So we are just interested in how these elements move in this space, but not how they biologically.
Do this so how they fly or how they see the space around and so on.
00:39:36 Dr Giulia Pisegna
What happens is that the huge extent of correlation that one can measure in space with the correlation length.
Reflection also in a logical relation in time, described by a quantity that we call the characteristic time scale.
The strong correlated physical systems are characterised by the fact that these two quantities, so the correlation length and their characteristic time scale and linked by a power law and bionic spoon, and that describes this relation.
As a result, it means that basically you need only the information about the value of this exponent to fully characterise the large scale dynamics of your system.
And one can prove that different physical systems with different microscopic details if they satisfy this expertise.
00:40:35 Dr Giulia Pisegna
They share the same value of the exponent and in this case, we say that they belong to the same universality class so surprisingly. They date about nature as worms showed that this system really verifies this property, but with a non trivial value of the Exponent.
That we think could really identify knew universality class for this kind of strong correlated biological system.
And in our recent paper, we therefore propose a study of a theory in order to reproduce the value of the exponent and the dynamical properties of the natural system of worms.
So we studied the model that had been previously introduced for flocks of birds, because these two systems, again apparently very far from each other, seem to share another important feature that is the mechanism of information propagation in the group.
That we are able to describe without specific set of dynamical laws. Properly defining a model.
And to understand if this was really useful for her biological system, we performed an analytic calculation on this model using the realisation with technique.
In order to relate the value of the exponent of this theory to the natural one.
An we obtain a very compelling and promising result that anyway needs lots of additional work to be a perfected but heavy hand of our analysis. We found a valuable exponent, almost consistent with the experiments.
Which confirms that our intuitions about the dynamical loss describing the system were correct. And also this calculation made has also to understand that this system of the size of a Metro system displaying collective behaviour is not just an arbitrary feature of this system.
Animal groups have to be large to have collected properties, but small enough to ensure the most efficient information proposition and to show the right physical quantities as the value of the exponent.
We also think that with this work we can say that there is no Magician Group is a theory that can be expanded and successfully use also for Heather application to biology.
The fact that we use the same model and mathematical equations for flocks of birds and images worms. And they also seem to work because they are really fitting experimental data.
Could be very interesting from the point of view to find a common explanation and description to the great variety of logical system. And so we really like this and we will like to go on along this line of research. In this concept lies the concept of universality. Also all the models that we use.
Must have an inspiration in the condensed matter physics in which one represents the items, for instance, of a particular element of matter, has just heralds or just spin that can align to other spins over other hatem’s for interaction. So basically we are trying to translate this simplification of theory. Also to the biological systems.
That in each individual you can recognise level of complexity because of course they are living systems, but if you look at the system far enough in a sense you can again described this. Particles has heroes that move in space and so trying to use the same models that you can use for condensed matter physics.
And back now with Professor Cavagna I can already see that there might be some risk of this research being either misinterpreted or misrepresented as having a formula for feet up crystal ball, or prediction through renormalization group research that will have an answer for everything, and I can imagine this is something that you might have concerns about it.
00:45:27 Prof Cavagna
Yes, of course, because on one hand there is the hope of using and applying exporting these fantastic tools. Statistical physics to biology, on the other hand, you have to be very careful.
So one hand you want to be able to forget about some of the details that make biology so **** ** the other hand, you have to remember that you must remember that.
A certain extent, biology is all about the details, so yes, it is not a miracle thing, so we need to be careful.
For example, we made experiments about Starling’s and we got some results.
Now Starling’s are just one species and Moreover. We did experiment just in one location here in Rome. So imagine that we now go for to another species of birds doing flocks like downlinks for example, there are beautiful flocks and balance.
Do we expect the collective behaviour of this system to be totally different?
Well, first of all. One should do the experiment and we don’t have the data on many different species of birds. We only have Starling’s, but.
My expectation, as a physicist would be, well, probably the collective behaviour that really the large scale behaviour is not that different. Of course, this is by now, simply hope both from despair, mental from the theoretical point of view.
But that in a way, is the very reason why I do this kind of studies. If I told that whatever I discover at the collective level in a certain system, a changes dramatically in a system be just because I’m no longer studying the stylings in Rome, but studying the starling’s in Milan.
Well, first of all, that is a motivation with me down a little bit . But also because when I do the experiment an I study the theories tying together this experiment, I really see that the microscopic details lose some of their relevance when I integrate up to larger scale.
00:47:48 Prof Cavagna
I’m not very optimistic to be able to classify biological system in so few classes as we do for physical systems. So certainly the great complexity and diversity of life Would reflect on stronger limitations of the applicability of the real meditation group in biology, okay?
And Moreover, one has never just trust that theoretical mechanisms that we develop, or one should always cheque this. So we have anyway to perform more experiment to compare our.
Theoretical prediction with the with the heart fruit.
And apart from the diversity in the huge complexity of life, there is another issue, which is that the physics of the randomization group was developed mainly in what we call equilibrium systems in physics, while biology is the quintessential out of equilibrium science. So there is a constant injection and dissipation of energy. So biological system are all.
All while the out of equilibrium.
So even though there had been some step forward in the physics of out of equilibrium systems, we can safely say that is much less under control then the physics of equilibrium systems.
So in this respect, I believe that not only we have to push forward the boundaries of the biology, but of physics set.
So all these challenges from the biology of collective system is actually stimulating a physicist to broaden the scope and capabilities of theoretical physics.
We are not simply applying the good old methods of statistical physics to biology, not at all, because they are not sophisticated enough. What we’re doing is that we are trying to apply the same concept.
But for doing that we have to expand the tools that we’ve been using up to now in in physics.
Now, to just briefly come back to something that you mentioned about Swarm behaviours earlier, I’ve just remembered seeing videos of Micro Copter display in place of fireworks oversold using little drone copters to fly around in shine messages out into the Sky. I was wondering if you had any thoughts on the application of renormalization group findings about swarm behaviour in engineering or an industrial ways like this or possibly coming at it from a different angle to take the population behaviour findings and coming up with the kind of social media software way of understanding or maybe even influencing behaviour.
00:50:55 Prof Cavagna
I truly believe that there should be room. There should be funds and there should be encouragement for science, which is not driven by applications but just…. Driven by curiosity.
And I think that this is actually good also for the technological advancement of humanity, because I think that most, if not of the machinery, is that we are using today for doing this interview. A computer microphone lies every.
Thing was originally developed by building on ideas of people who were totally completely uninterested in the applications of their ideas.
All of these, I mean you. You simply can think about the determination principle by then Eisenberger that was something incredibly abstract and theoretical, which ended up funding the entire quantum mechanics and all day digital devices that were using today.
So yes, I can answer about that is not something. However, that motivates might research, however this said.
There is in fact a huge technological interest for the for the research on collective behaviour in biology and the interest is quite obvious. Is that of distributed behaviour so.
Yes, there are artificial networks of everything of drones of devices of all swords, and they will increasingly more in the future.
The problem is how to control these swarms. Yes, you can call them artificial Swans now. The classic way of controlling that is a kind of top down control in which you have some central control that’s control all the peripheral arms of the system.
This kind of control is kind of easy to implement, but has a problem to be quite fragile because if you go and you damage in some way the central control you have a problem with the whole system.
So This is why people are increasingly turning to nature to get inspiration, because in collective biological system typically you have a distributed control, so sort of bottom up control.
In the system we study in Starling, flocks swarms and so on. You do not have any head of the system any.
Leader, you may have leaders in other kind of behaviour, like when birds are migrating from one location to another, but in the incredible display done by Starling’s in the evening during winter with absolutely no central leader coordinating them. So that is a sort of distributed.
Control, which is clearly much more robust because no matter how many units get damaged, you still get very, very stronger coordination of the SIS.
Now the possible application of these are many from the nice and desirable to the less desirable of military nature.
Oh yes, there is a great technological interested in that I’m not sure how much the RG the realisation group will contribute to that technological advancement. My concern is only to try and find the most general lowest possible to quantitatively describe collected biological systems.