(highlight:: The Structure of a Network Differs Based on the Type of Information Being Passed Through It
Summary:
We often think about different occurrences on a network, but they don't necessarily have the same structure.
This contrasts with the prevalent view in the literature. Using different methodologies can reveal different perspectives and understandings of a concept, like studying the economy algebraically versus algorithmically.
It is valuable for people to approach problems in different ways and not assume that one approach is superior for everyone.
Transcript:
Speaker 1
But that's the sort of perspective that I think is interesting, because even though we can do everything as a network, and there is a notion of network structure, usually we're thinking About something occurring on that network, and I don't see why it should be the same structure that shows up if it's a different thing occurring. I don't think that has to be true.
Speaker 2
But that's not the prevalent view of this in the literature. Interesting, because I feel like I oversight actually at this point on the show, the conversation I had with Brian Arthur a while back, 68 and 69, where he was talking about the way that Different methodologies disclose or enact different ontologies. And specifically talking about the way that if you're using algebraic thinking to study the economy, you're going to get a whole different view of what an economy is than if you're thinking Of it algorithmically. And so this is an object or process thing. Sure.
Speaker 1
I definitely think it's good for people to study problems in different ways. I think it's important for people to use different approaches, and to not just say, oh, this approach is the best everyone else should do it that way too. I think that's okay.)
- Time 0:18:12
- pluralism, network structure, network topology,

(highlight:: Conveying the Portfolio of Techniques in Complex System Science at SFI Institutionally
Transcript:
Speaker 2
But then there's this other thing, which I think as a way of conveying what seems like it makes the status quo of complex system science as practiced by SFI institutionally is that people here seem to enjoy bringing to bear a whole different portfolio of techniques on a single question and not assuming that one is going to outperform the rest. Yeah. So at any rate, that having been said, I wanted to double back because in as much as this is me just getting to indulge my own noob learning in public here, I want to talk a little bit more about the way that these networks, these techniques reveal modularity and hierarchy in communities. And I'd love to hear you riff on that because like there's, this is just to call this shot. This is a way of setting the table for later in this conversation, talking about paper that you led on the topology of data and how different layers of granularity or like fuzziness, different levels of resolution of data seem to reveal. Again, like completely different, like different objects, like the jiggly puff thing, but I don't want to- Right. Right.
Speaker 1
Yeah. But so it just- Yeah. And we published a Pokemon in a serious physics venue may not happen again. We got away with it, which is good. Jiggly puff is, by the way, like if you've seen the detective Pikachu, indignant jiggly puff is my spirit animal, but specifically the indignant type.)
- Time 0:20:56
-

(highlight:: The Level/Scope/Depth at Which You Look at a Phenomenon Significant Influences How You Understand It
Summary:
The text discusses the intersection of methodological pluralism and the impact of studying a phenomenon at different levels.
It mentions individuals like David, Jess, and Nihat who are researching what defines an individual. This inquiry is closely tied to the question of community detection.
Transcript:
Speaker 2
So for me, it's like on the one hand, you have this sort of methodological pluralism, but then on the other hand, you have all of this work that different people have done in the SFI network On how the granularity or the level at which you are studying a phenomenon is really going to radically change the way that you understand it. And like, we've got like David and Jess and Nihat and these others that are working on what exactly constitutes an individual. And I just find this question about community detection very closely related to that kind of fundamental inquiry.)
- Time 0:22:28
-

The Difference Scales of Community Network Detection (Micro, Meso, and Macroscale
Summary:
Community detection involves analyzing mesoscale structures which are neither micro scale nor macro scale.
Communities are a specific type of mesoscale structure characterized by dense internal connections and sparse external connections. Other mesoscale structures include core periphery structures and role structures.
While core periphery structures also have dense connections, they are connected to other things in a non-sparse manner.
Role structures focus on identifying similarity in interaction patterns rather than literal density.
Community structure is the most studied type of mesoscale structure, but advancements have been made in other types as well.
Transcript:
Speaker 1
I view community detection as something that we're doing to look at a mathematical object and it then gives an output. And the type of structure we're looking at is a specific type of what I would call a mesoscale structure. So hopefully this is starting to get at what you're bringing up where the micro scale structure might be the individual nodes or potentially individual edges and so on. And a macro scale structure might be distributions of stuff, whether it's distributions of the number of friends or whatnot, do you have many people with few friends or lots of or a few People with many friends. So it's macro scale in that sense. The mesoscale is stuff in the middle and communities are a particular type of mesoscale structure. In particular a type in which you have dense connections inside, whatever that means, rinse sparse connections between. There are other types of mesoscale structures, like there's something called a core periphery structure where you still might have dense stuff, but it's connected to other things, And not necessarily in a sparse way. There is something that social scientists call roles and positions or like role structure, where it might not be that you have literal denseness, but maybe the interaction pattern Say, oh, a faculty member has a certain type of interaction patterns, like a local network structure, and a graduate student has a certain other type of interaction pattern, another Network structure presumably with more friends than the faculty member has. But you're identifying similarity of the structure rather than literal denseness. So there's different types of mesoscale structures that people study, and community structure is somehow made. So one that people have advanced the most on, I would say, is the most straightforward to think about it. There's advancements on the others too.)
- Time 0:23:22
- network structure,

(highlight:: Representing Social Interactions/Relationships as Edges in a Network Is A Poor Approximation of Reality
Summary:
Interactions between people are much more complex than a simple mathematical representation.
We can't accurately capture the nuances and multiple types of relationships in a matrix or network. When we study these mathematical objects, we must be cautious about the simplifications we make and the potential distortions that occur.
Our goal is to derive meaningful insights about the real world from these mathematical representations, but we must remember that they are a simplified version of reality.
Transcript:
Speaker 1
The reality is if you look at this, it's like, okay, yeah, as a first pass approximation, whether somebody or except your friend request or whatever, but like there are so many different Pieces to this. So we do it because we do a lot of these things because we know how to do them, not because we necessarily think it's a beyond and all. And if you're going to go beyond pairwise interactions or if you're going to go to have interactions that change with time or if you're going to go to have interactions that have maybe Multiple types of interactions, multiple communication channels, multiple types of relationships, you're going to end up having a mathematical representation that's more complicated. And I've spent a lot of time studying various ones of those and other people spent a lot of time studying various ones of those. And it's a very cool thing to do. And there's good reason to do it because in the way that I probably phrased in my talk because I often phrased it that way, people are not walking around holding sticks that other people Are holding. Right. Even the notion of saying two people are friends, there is this latent set of interactions between them, phone calls and emails and going to the movie and going out to eat and what not That you're not seeing and you are representing as a number in a matrix saying that you're not seeing it, these two people are holding an attached to a stick and we're using a mathematical Representation that says people holding a set of stick and that's not what we see. Right. What we see is two people are having dinner together or whatever. Right. Well, hopefully you don't see that because you shouldn't be stalking them but interactions and somehow we represent the interactions as a network. Yeah. I'm very snarky and make comments like this all the time, apologies. But in any event, so we're representing something in a mathematical object that even when we include a bunch of these nuances, non-parwhites interactions and so on is a much simpler Thing than reality. It's a gross simplification of reality and you always have to worry that when you get reality and then you get data from reality, which is already making certain simplifications and That you then turn it into a mathematical object that you study, there's this thing of, okay, if I turn it into a mathematical object, I can say that potentially I am making a precise statement About the mathematical object, although even then there's approximations, but let's suppose that I'm doing that. Now I've made a precise statement about mathematical object and I want to turn that statement and imply something and say something about the real world, even though the mathematical Object is a simplification of the real world and you have to worry because there are artifactual things that occur by choosing to have represented something in a certain way and the Hope is that something that you then say about this object hopefully can tell you also something about the more complicated thing it's representing.)
- Time 0:41:15
-

(highlight:: The Structure of a Network Differs Based on the Type of Information Being Passed Through It
Summary:
We often think about different occurrences on a network, but they don't necessarily have the same structure.
This contrasts with the prevalent view in the literature. Using different methodologies can reveal different perspectives and understandings of a concept, like studying the economy algebraically versus algorithmically.
It is valuable for people to approach problems in different ways and not assume that one approach is superior for everyone.
Transcript:
Speaker 1
But that's the sort of perspective that I think is interesting, because even though we can do everything as a network, and there is a notion of network structure, usually we're thinking About something occurring on that network, and I don't see why it should be the same structure that shows up if it's a different thing occurring. I don't think that has to be true.
Speaker 2
But that's not the prevalent view of this in the literature. Interesting, because I feel like I oversight actually at this point on the show, the conversation I had with Brian Arthur a while back, 68 and 69, where he was talking about the way that Different methodologies disclose or enact different ontologies. And specifically talking about the way that if you're using algebraic thinking to study the economy, you're going to get a whole different view of what an economy is than if you're thinking Of it algorithmically. And so this is an object or process thing. Sure.
Speaker 1
I definitely think it's good for people to study problems in different ways. I think it's important for people to use different approaches, and to not just say, oh, this approach is the best everyone else should do it that way too. I think that's okay.)
- Time 0:18:12
- pluralism, network structure, network topology,

(highlight:: Conveying the Portfolio of Techniques in Complex System Science at SFI Institutionally
Transcript:
Speaker 2
But then there's this other thing, which I think as a way of conveying what seems like it makes the status quo of complex system science as practiced by SFI institutionally is that people here seem to enjoy bringing to bear a whole different portfolio of techniques on a single question and not assuming that one is going to outperform the rest. Yeah. So at any rate, that having been said, I wanted to double back because in as much as this is me just getting to indulge my own noob learning in public here, I want to talk a little bit more about the way that these networks, these techniques reveal modularity and hierarchy in communities. And I'd love to hear you riff on that because like there's, this is just to call this shot. This is a way of setting the table for later in this conversation, talking about paper that you led on the topology of data and how different layers of granularity or like fuzziness, different levels of resolution of data seem to reveal. Again, like completely different, like different objects, like the jiggly puff thing, but I don't want to- Right. Right.
Speaker 1
Yeah. But so it just- Yeah. And we published a Pokemon in a serious physics venue may not happen again. We got away with it, which is good. Jiggly puff is, by the way, like if you've seen the detective Pikachu, indignant jiggly puff is my spirit animal, but specifically the indignant type.)
- Time 0:20:56
-

(highlight:: The Level/Scope/Depth at Which You Look at a Phenomenon Significant Influences How You Understand It
Summary:
The text discusses the intersection of methodological pluralism and the impact of studying a phenomenon at different levels.
It mentions individuals like David, Jess, and Nihat who are researching what defines an individual. This inquiry is closely tied to the question of community detection.
Transcript:
Speaker 2
So for me, it's like on the one hand, you have this sort of methodological pluralism, but then on the other hand, you have all of this work that different people have done in the SFI network On how the granularity or the level at which you are studying a phenomenon is really going to radically change the way that you understand it. And like, we've got like David and Jess and Nihat and these others that are working on what exactly constitutes an individual. And I just find this question about community detection very closely related to that kind of fundamental inquiry.)
- Time 0:22:28
-

The Difference Scales of Community Network Detection (Micro, Meso, and Macroscale
Summary:
Community detection involves analyzing mesoscale structures which are neither micro scale nor macro scale.
Communities are a specific type of mesoscale structure characterized by dense internal connections and sparse external connections. Other mesoscale structures include core periphery structures and role structures.
While core periphery structures also have dense connections, they are connected to other things in a non-sparse manner.
Role structures focus on identifying similarity in interaction patterns rather than literal density.
Community structure is the most studied type of mesoscale structure, but advancements have been made in other types as well.
Transcript:
Speaker 1
I view community detection as something that we're doing to look at a mathematical object and it then gives an output. And the type of structure we're looking at is a specific type of what I would call a mesoscale structure. So hopefully this is starting to get at what you're bringing up where the micro scale structure might be the individual nodes or potentially individual edges and so on. And a macro scale structure might be distributions of stuff, whether it's distributions of the number of friends or whatnot, do you have many people with few friends or lots of or a few People with many friends. So it's macro scale in that sense. The mesoscale is stuff in the middle and communities are a particular type of mesoscale structure. In particular a type in which you have dense connections inside, whatever that means, rinse sparse connections between. There are other types of mesoscale structures, like there's something called a core periphery structure where you still might have dense stuff, but it's connected to other things, And not necessarily in a sparse way. There is something that social scientists call roles and positions or like role structure, where it might not be that you have literal denseness, but maybe the interaction pattern Say, oh, a faculty member has a certain type of interaction patterns, like a local network structure, and a graduate student has a certain other type of interaction pattern, another Network structure presumably with more friends than the faculty member has. But you're identifying similarity of the structure rather than literal denseness. So there's different types of mesoscale structures that people study, and community structure is somehow made. So one that people have advanced the most on, I would say, is the most straightforward to think about it. There's advancements on the others too.)
- Time 0:23:22
- network structure,

(highlight:: Representing Social Interactions/Relationships as Edges in a Network Is A Poor Approximation of Reality
Summary:
Interactions between people are much more complex than a simple mathematical representation.
We can't accurately capture the nuances and multiple types of relationships in a matrix or network. When we study these mathematical objects, we must be cautious about the simplifications we make and the potential distortions that occur.
Our goal is to derive meaningful insights about the real world from these mathematical representations, but we must remember that they are a simplified version of reality.
Transcript:
Speaker 1
The reality is if you look at this, it's like, okay, yeah, as a first pass approximation, whether somebody or except your friend request or whatever, but like there are so many different Pieces to this. So we do it because we do a lot of these things because we know how to do them, not because we necessarily think it's a beyond and all. And if you're going to go beyond pairwise interactions or if you're going to go to have interactions that change with time or if you're going to go to have interactions that have maybe Multiple types of interactions, multiple communication channels, multiple types of relationships, you're going to end up having a mathematical representation that's more complicated. And I've spent a lot of time studying various ones of those and other people spent a lot of time studying various ones of those. And it's a very cool thing to do. And there's good reason to do it because in the way that I probably phrased in my talk because I often phrased it that way, people are not walking around holding sticks that other people Are holding. Right. Even the notion of saying two people are friends, there is this latent set of interactions between them, phone calls and emails and going to the movie and going out to eat and what not That you're not seeing and you are representing as a number in a matrix saying that you're not seeing it, these two people are holding an attached to a stick and we're using a mathematical Representation that says people holding a set of stick and that's not what we see. Right. What we see is two people are having dinner together or whatever. Right. Well, hopefully you don't see that because you shouldn't be stalking them but interactions and somehow we represent the interactions as a network. Yeah. I'm very snarky and make comments like this all the time, apologies. But in any event, so we're representing something in a mathematical object that even when we include a bunch of these nuances, non-parwhites interactions and so on is a much simpler Thing than reality. It's a gross simplification of reality and you always have to worry that when you get reality and then you get data from reality, which is already making certain simplifications and That you then turn it into a mathematical object that you study, there's this thing of, okay, if I turn it into a mathematical object, I can say that potentially I am making a precise statement About the mathematical object, although even then there's approximations, but let's suppose that I'm doing that. Now I've made a precise statement about mathematical object and I want to turn that statement and imply something and say something about the real world, even though the mathematical Object is a simplification of the real world and you have to worry because there are artifactual things that occur by choosing to have represented something in a certain way and the Hope is that something that you then say about this object hopefully can tell you also something about the more complicated thing it's representing.)
- Time 0:41:15
-