 Hello, everyone. Welcome to the active inference live stream. It is active stream 9.2 and it is December 1st, 2020. So happy December to everyone. Today we are really excited to have David, one of the authors of the paper we're going to be discussing. So it will be just a great discussion today. Really looking forward to it. Welcome to TeamCom, everyone. TeamCom is an experiment in online team communication, learning, and practice related to active inference. You can find us on our website, at Twitter, via email, at our YouTube channel, or our public keybase team and keybase username. This is a recorded and an archived live stream. So please provide us with feedback so that we can improve on our work. Also, all perspectives and backgrounds are welcome here. And as far as video etiquette for live stream goes, mute if there's noise in your background, raise your hand so that we can hear from everybody in the conversation and let's use respectful speech behavior. Just some quick updates on upcoming active streams. Today we are in 9.2 December 1st talking about this project of consciousness model paper. And in the rest of December, we'll be talking about two papers on December 8th and 15th will be in active 10 on a variational approach to scripts. And on December 22nd and 29th will be in active 11 on sophisticated effective inference. So to cool and very interesting papers that we'll be getting to. But today in active nine, we're still continuing our discussion of project of consciousness model. First, we're going to have some introductions and warm ups. Then we'll go through 9.2 as with previous times, the project of consciousness model paper by will afford at all 2018. And in last conversation in 9.1 and 9.0, we talked about some of the groundwork and the aims and claims of the paper as well as going through the abstract and the roadmap. Today, especially because we're lucky enough to have an author on board, we're going to be talking about a few deeper and auxiliary questions, as well as going through a few key quotations and talking about maybe some of the implications for the paper. And then as always, we'll return to the figures and ask what each figure is representing or invoking. So cool to have this discussion today and then different papers in 10 and 11 and then we'll have a whole new suite of papers in 2021. So get in touch with us if you want to participate. First though, with our nice small group today. Let's introduce ourselves. We'll go around and give a short introduction or a check in and then just pass it to somebody who has not spoken yet. So I'm Daniel, I'm in California, and I will pass it to Adam. Hi, I'm Adam, Adam Salfron. I'm in Bloomington, Indiana at Indiana University as a post doctoral researcher. And I'm very interested in consciousness and recently came up with a theory I called integrated world modeling theory, which has a lot of overlap with David's work and which I find to be really fascinating and inspiring. Glad to be here. Cool, let's go to David. Hello there, I'm David Rudroff. I'm an associate professor at the University of Geneva, Switzerland. And I'm one of the co-authors of this paper and somebody who work on trying to model the role of consciousness, including its phenomenal aspects. So in that way, generally speaking, biological cybernetics and more particularly active in France. Thank you for having me here today. Cool, let's go to Sasha. Welcome, Sasha. Hi, everybody. My name is Sasha. I'm in Davis, California, and I study developmental neuroscience and how synapses form in the developing brain. And I'm really interested to be part of this conversation today. Cool, then we'll go Alex and Stephen. Hi, everybody. I'm Alex. I'm in Moscow, Russia, and I'm a researcher in systems management school. And I visit to Steven. Hello, I'm Steven. I'm based in Toronto, and I'm doing a practice based PhD in to participatory theater approaches and the use of landscapes to explore themes. And back to Daniel. That introduces our nice group today. As far as the warm up questions go, we'll just have people raise their hand whenever they feel like it. And I'll just drop out the first two questions at the beginning, and they are what interests you about studying consciousness and what is a way that we can study it. So both David and Adam, I know you've thought a lot about this. Maybe one of you two would want to go first here. Yeah, David, just how did you get to this area? It's just a long story, but to make it short, let's put it that way. It's really related to philosophical and existential question and the quest for turning this type of question into something scientific. But actually, I was not that interested in psychology when I was younger. I wanted to be either an astrophysicist or a conductor of orchestra, but I did not. Interesting. It's a different type of coordination that you're achieving now. Steven, and then anyone else who wants to raise their hand. Yeah, I'm interested in consciousness. In a way, it's kind of a late arrival because I always thought that was something esoteric that I couldn't really participate in. And I've suddenly seen this as starting to come centre stage in both this active inference world, but also being talked about more in the participatory theatre environment. So what does it mean to be consciously aware, to have imagination? And what I think is quite interesting is to try and move beyond this kind of pan-psychic approach that seems to permeate a lot of people who actually go there. It's either it's something that's ignored and pushed to the side, or it's everywhere and everything in this pan-psychic. And what I'm really interested in is how can it be engaged in terms of how we are reflective or reflexive or participatory, and what consciousness means in terms of talking to people who aren't the same as us. Interesting. Definitely the coming from a different perspective is something that's baked into this type of a framing. Because, yeah, everyone's coming from a different perspective. That's our starting point. That's not our conclusion here, and then there's questions about how we're going to coordinate amidst that. So we can either have anyone else raise their hand on these questions, or I'll just put up this special point to question, which is what is something that you have thought about differently after reading the paper or participating in last week's discussion? And for me, it was actually biking in the morning and seeing the moon just look massive and thinking about that illusion in the context of the projection and what we talked about last week, as well as some other literature by these authors. So I just thought what might be some differentiable experiences, some little tests that we could do that would be able to capture differences between the projective model and other models of consciousness. I didn't really have an answer, but it was just something I was thinking about after our discussion. Anyone else after reading the paper or participating in the discussion? Stephen? Well, one thing that it was really helpful this unpacking the diagrams last week. And I noticed that there was that box for the head of the perspective, the consciousness, the perspectival consciousness. And it suddenly made me realize that, yes, this can bridge more back into the embodied work, because I'd had a discussion about this paper before with someone who's in theater and interested in kind of some of this more kind of esoterical kind of ways of knowing. But I'm trying to find a way to bridge it back to the embodied sort of inactive work. And I sort of, by seeing it in the kind of head, I suddenly realized, OK, it still does come back to the physicality. So that was really, really useful for me. Cool. Cool. Sasha? Yeah, after last week, I just began seeing everything as, you know, more geometrical and thinking about people and their behavior as projections, kind of, yeah, recognizing that we're always seeing just one side or one perspective of someone's behavior. And specifically that that GEB logo, it just, that's all I could see. Right, with the light from the different side showing the different letters, GEB. Yeah. And we'll come back to GEB a little bit later in this discussion. Any other thoughts from the panelists on these warm-up questions? Otherwise, I'm sure there'll be a lot to get into, a lot of tangents to go off onto. Cool. All right. David, go ahead. No, I just wanted to say that there would be a lot to say about the points and questions that were raised, but perhaps we can not do that in power and go through the stream and see your questions come about. Great. So today in this discussion, as with our last two weeks, we've been talking about the Projective Consciousness Model and Phenomenal Selfhood, which is a paper in Frontiers in Psychology, December 2018 by Willa Ford et al. And thanks, David, again, for coming out to the stream. It's always awesome to have the authors. The aims and claims of the paper we did go through in a previous week, but David, just I'm curious here, how do you present this work or how do you introduce this topic to different audiences who might have familiarity with different topics? Like, how do you frame the aims and the claims of the paper? So, lately I've been particularly talking in contexts where people come from artificial intelligence, robotics, virtual reality with, you know, of course, more broader philosophical interest. And I generally try to frame it starting a bit like you did. I mean, from the question of, you know, what we could call biological cybernetics and you elaborated on that term before and how active inference for situated autonomous system, you know, with survival challenges is an interesting padding to understand cognition within biological cybernetics. And then I come to the question of what would be then the role in that process of what, you know, we intuit as being consciousness what we can say about it, emphasizing the big and important challenge of connecting, let's put it that way, for short. You know, the phenomenology of subjective experience is some aspect of it with, for lack of a better word, functions. Why do we perceive and imagine the world the way and others and yourself in the world the way we seem to do at a certain level and how that actually, how can we make sense of that to understand computational in some sense process that are key to, you know, adaptation, you know, resilience in the context of biological cybernetics. Cool. Thanks for that. And I think it was really clear how you laid out that connection in the paper. Basically, there's the functions of consciousness, which we'll go through soon. And then there's the subjective aspects, which were captured in the invariance. And the PCM is like a bridge that's going to be connecting the subjective invariance of consciousness to some of these functional attributes. And so we're kind of going from point A to point B across a specific river. And so if we can really be clear about which sites on the riverbank we're building from and then be really clear about what the bridge is and isn't, then we'll have a really good discussion about consciousness as opposed to just swirling around and thinking it does everything and nothing is everything and nothing at once and all these things. So any thoughts on the aims and claims from anyone else? Otherwise, we'll just go through to the next section. Cool. Yep. We have a lot of questions coming up. So in previous weeks, we looked through the abstract sentence by sentence just to understand how the authors were taking the free energy minimizing approach of active inference, biological cybernetics and then connecting this to a few other areas such as geometry and the embedded psychology and that was the abstract. In the roadmap, we had the sort of outline of where they go. They start with an introduction and go straight to the methodology and then talk about the two river banks in that previous metaphor that we want to connect the phenomenological invariance of experience and then the functional features of consciousness and then sets up the problem in a way that hopefully could be solved. So maybe David, I'm curious like in other papers, how have you structured it differently or this was in 2018? What have you been like wishing that could have been explained more or where have you bolstered the theory since then? All right. That's a huge question. So number one, Kenneth Williford, the first author of this paper we expected him to join. I don't know if he will join later. He's a friend of mine and I admire for the quality of his mind. He's a philosopher. And so the structure of this particular paper we look at today is I think deeply informed by his philosophical methods. If you were to read the first paper we published on the topics in the Journal of Biological Theoretical Biology, sorry, in 2017 you would find a two-worldy though but very different way of presenting but they are totally consistent with each other. And in both cases, you know, it's still working progress so there are still a lot of questions and of course we are not claiming that we explain everything by far. We're trying to restrict ourselves to things we think we can frame in a classical scientific framework. And, you know, I've been working for years now in my lab and we are actually finishing the paper on that to integrate in this, in the general, based on the general principle of this model in addition to visual illusion as we did recently in the Moon Illusion paper that was published in 2020, mechanism of affective and social appraisal and processing and social perspective taking based on projective geometry not only but not only to model and understand how artificial agents, virtual or robotics as a model of ourselves if you wish can do theory of mind and how that plays a role in emotion regulation, adaptation of behavior and that's basically what I've been doing and interestingly, without taking too long, there are some studies from psychophysics that show or that, you know, based on empirical data that if you are confronted with an object or something you feel is dangerous like, for instance, a snake, for some phobic, for instance, the intensity of your fear will be a function of the inverse, more or less the inverse of the distance of the object and that has been formalized as being, you know, something you can identify with an empirical psychophysical law called the Stevens law with a certain type of exponent and as it turns out, actually, the projective geometrical framing just gives you that, naturally, it transforms information the size, the apparent size of objects according under certain constraints to a law that is an inverse from the distance, more or less so you can recover things that have been documented before and understand how they may play a role in appraisal something I feel when it's closer, it's bigger, it takes more of my field of consciousness if you wish and so one way of motivating my behavior of, for instance, avoidance is imagining myself farther so the thing is smaller, takes less space and as a result it has less impact following the very same law we're discussing and therefore that will minimize my free energy and I will move away and I do that with robots and things like that Very cool, yeah, I'm imagining this, like, fear is proportional to one over distance something in that family of equations like a bus that's 100 feet away, not scary 50 feet away, not scary it's still on the flat part of the fear curve and then it gets exponentially higher when it gets closer to you and takes up more of your field of vision and then there's also long distance threats somebody with a projectile or something like that and so that entails a causal model where somebody who's doing a certain action will be able to make an object quite big in your face very soon but maybe it's not quite there yet but it could happen rapidly so that's kind of an interesting take Yeah, so the point here is along the line of what they were saying trying to understand what could be some aspect of the causal role of consciousness in things like appraisal and motivation of behavior here the point that we make is that actually the spatial structuring that we put forth based on projective geometry implies naturally something about relative size that is consistent with psychophysics and so it's to say that somehow this structuring of space by consciousness act as a sort of lens, a magnifying device if you think about it that way that in virtue of its own geometry will have an impact that can then be used for other aspects of motivation, programming and behavior in the context of active inference So maybe we'll get back to this intermodal area again but that's the visual field and not all organisms or humans may have the visual field but it's clear that the spatial objects we visualize them in space and there's something similar though like if there's a lot of noise in the background but something is what you're taking up your field of audio awareness and so it's like in each of our senses there can be like a sensory expansion of focus Greetings, Lee, just play on we're just continuing the discussion but go ahead there, David Yes, so one of the essential things we emphasize is that it's actually not just the visual field it is the cognition of space as in form in a multimodal manner so in audio you may recognize objects and all of a sudden kind of an image of sort the spatial representation of that object as if you could see it even though you don't see it blind people do that all the time they have actually perspective representation in the represent object in space based on other sensory channels there is an inference that is being made and the idea is that consciousness operates at some level of abstractions so you have actually extracted information from different sense modalities and then they become spatial objects if you wish with variety of properties that you put in perspective in that field of consciousness and based on which you will operate at the conscious level if that makes any sense Yep, interesting it's like it's so hard sometimes to disentangle from the eyes, vision, perception debacle, nightmare but we're not experiencing what photons hit our retina we know it's a generative model and you're really taking it to the next level you're not even just seeing a generative model of the visual field you're seeing this extremely your quote seeing this integrated field of consciousness and that may depend on your visual abilities or your acuity but it also something that's quite derived from intermodal stimuli and yeah, well, that's cool stuff let's get to this projective geometry idea and we as non geometers talked about it a little bit in previous discussions and just try to capture a few of these points that you said were differing critically between Euclidean and non Euclidean geometries so I'm just curious and perhaps those who are not familiar with hearing about geometry and consciousness in the same sentence just what is it that brings geometry to the table and how do we end up even talking about geometry and consciousness together Excellent question, so first of all a little disclaimer I'm not a mathematician even though since I understood early on that's what I was trying to do required to think mathematically fairly advanced way I banged my head to try to progress but I'm privileged to work with real and very good mathematicians so that helps but nevertheless I can give you to the limit of my knowledge some info about it number one I would say generally speaking if experience one conspicuous aspect of experience of subjective experience is a spatial manifestation not all experience reduced to space but a big chunk of it is involved so it's good from a scientific standpoint because then it becomes something you can think about mathematically and a lot of models in let's put it that way incognitive science has been based on information processing sometimes very formal model like advanced Bayesian models or even active in France and so there is a lot of emphasis on the mathematics of information theory so to speak and here the notion is that in order to breach information processing if you wish with aspect of experience as manifested as a space geometry comes obviously as a scientific object as a mathematical object as something necessary so the challenge is then to understand how to articulate information theory with geometry in order to have a fuller account of the process so then now on the specifics of projective geometry it starts from trying to find some let's call it that way phenomenological invariant that most people would agree upon to a certain extent considering experience introspectively it appears that it's obvious in vision but they claim it's general most perception and imagination that the way we quote-unquote represent the world let's not enter into the metaphysics of representation let's use it in a pragmatic way here this notion of representation is in perspective obviously parallel line do not appear to us in vision as parallel but they converge at a certain vanishing point toward infinity there is an horizon and it's three-dimensional it's not like just a projection of a 3D world of an image we experience it with a sense of depth and I argue that the same when you imagine yourself somewhere considering something aiming at something the notion of aiming at is very important in the phenomenological tradition in philosophy it's not just about vision it's about spatial aiming in all those circumstances the world is mapped either a world that seems to correspond to our surrounding a world that corresponds to things we remember or even a totally fantasy world of things that may never exist every single time we build our experience in consciousness the point is that there is a perspective that is being taken with these properties of having an elusive point of view in perception we tend to feel it we feel that we are in the middle of that space so that the origin is somewhere in your head perhaps but that's not necessarily the case and that's not something that we can identify as the way we could identify the location of an object in space it's elusive, something that the phenomenologists have talked about a lot and it's in perspective so if you take those two things the origin of the projection or of the point of view that is elusive and a three-dimensional representation of space that is in perspective then the only geometry the D geometry that account for that is projective geometry and here since we say the space is in three dimension there is this notion of left, right, up, down and forward and backward it's a three-dimensional projective space and then that means that this space is ruled by the so-called projective transformation that are specific type of transformation that comes from a group of transformation that transform perspective into that space so for instance you can transform a point of view from the point of view I have here looking at that screen to an imaginary point of view I would have in a different position myself or somebody else naturally governed by that geometry so then that geometry appears naturally as A or if not D candidate to account for this aspect of phenomenological experience and therefore then you can start thinking about how that can play a role in the process of active influence nice thank you for the clear explanation really in that little bit the clearest part moving forward is this elusive origin spatial perspective that we have if someone can imagine an experience that doesn't fulfill those two conditions then we're already breaking our assumptions but if you can that's great it's like write a paper about it but given that there's consensus that there's an elusive origin and a spatial perspective we can then say that we're in a family of geometries that are related to this projective transformation and so going from there there's other downstream consequences to this whole topic sometimes to talk about of course transdisciplinary ideas but I think it's really clear how the structure of your argument is set up Steven just a couple of questions is the distance as things go away is that and you feel less intensity in terms of the fear or something is that linear or is that kind of exponential or is it very dependent on the space and I suppose the other question is is this transformation between one mapping to the kind of projective geometry so it's not like you store it in that form there's some sort of action to make it into this projection yes very good question I'm going to try to address them to the extent I can so number one these experiments that shows the spatial relationship between subjective fear and distance they found is a power law with an expression minus one so basically it's like the fear decrease as a function of the inverse one over the distance so it's not linear it's not exactly an exponential it's not linear and as it turns out that's the very basic rule of perspective when you want to build a space in perspective you divide by your coordinates XYZ coordinates by a parameter that corresponds to the distance along with you are basically projecting the direction of aiming that can be for instance in front of you so basically you find that in that geometry there is a similar operations if you wish as those that are implied by the one over distance models that is empirically derived so that's non-linear but it's a very particular type of non-minority that is not a super complex non-minority that can be handled computationally which I guess is important and now for the second part of your question can you remind me of it was just about that transformation between however the brain stores it and this projective phenomenon that we experience yes so the idea here somewhat what we have to presuppose if we want to embrace that type of model is that there is an unconscious world model which is complex which is constantly updated that somehow store information more or less in memory in an Euclidean manner and that we access that information by taking perspective on it so by transforming this sort of quote-unquote Euclidean representation into a perspective objective representation that involve multiple steps and that's something for instance we've developed the most basic one I would say in our morning vision paper in which we show and we how the calibration of a projective frame in 3D so we set the parameter that defines the projection and free energy minimization that's the idea free energy minimization acting on geometry entails illusions such as the illusion explain the illusions such as aims or illusions and many other regions we've not developed in this paper and that is somewhat the idea is that sometimes people talk about access consciousness let's not enter into the complexity of that debate but let's say that the hypothesis here is that in the way we access information that's in consciousness there is somewhat a change of coordinates quote-unquote from an Euclidean representation to a projective one and that operation is essential in what makes information conscious versus not, you know, model if that makes any sense thanks for the answer let's go to Lee oh hiya, so I was just wondering to build on Steven's question if a fear response is exponentially related to distance does that hold or is there any research it shows whether that holds for a counterfactual version so I'm thinking here about phobia or trauma so if you have a particular experience, you know, often in PTSD those things are replayed so yeah, so did you have any thoughts on that or any pointers? That's really a great question okay, the literature on this on the basic experiment is not huge surprisingly and your question is super interesting indeed because we would like to demonstrate that even if you do it in an imaginary way, you recover this type of effect consistent with the model I'm not aware of such study but I'm doing so much multidisciplinary work that the literature I have to cover is vast and sometime I may miss important studies in specific subfield but definitely that's that gives us the roadmap for a great experiment in some of the literature I'm thinking of Benjamin Bergen on simulation he talks about the difference between canonical perspective and I can't remember the other one but it's essentially first person perspective and the ability to shift between these so I just wondered whether it might be related but I'd be certainly interested if there's any research that you're thinking about doing yeah, David unmute and then go ahead and then Sasha so you want me to speak now if you have anything to add sorry I misunderstood I lost the screen oh yeah, no great point and thanks for that question Sasha yeah, very interesting stuff makes me think about this consciousness and understanding of how to interact with your space as like narrative building and how that develops over the lifetime that like as a child that the things that you're afraid of are quite specific and nearby and as you know as we grow up we're afraid of much kind of larger but more distant things but there's something some transformation we have to figure out as far as we still have to be motivated and afraid of the things that are either very afraid or very sorry very scary or very unlikely and those are still enough to motivate people to avoid them and so I think that's a really interesting something in there about changing the scope of your non Euclidean projection let's say as you grow up nice Sasha and it's kind of like when you go back to see something scary in childhood you go oh wait it was only a six inch stair that we were jumping off of well you're bigger but also you have a different worldview and then just like in the computational psychiatry realm you might talk about difference in a statistical parameter leading to decreased motivation or decision making it could be a distortion of this projection that results in things being perceived as dangerously close claustrophobia people are too close to me or people are too far away from me there's so much there with a social touch as well let's go just onto the sort of new parts of 9.2 and talk about some of these other slides and questions we have so last week we talked about the functional features the cybernetic features as well as some of these other theories threads that were drawn in like the availability kind of like a workspace the simulation theory motivation of action we talked about that last week as well as the questions about the phenomenological invariance with the emphasis on point one being that relational phenomenal intentionality just pause the video and look at those if you want more detail or of course read the paper but let's go to some feedback from participants so one thing that a participant who couldn't be here today asked was one thing I'm curious about is whether they are working on more complex forms of the model with higher temporal depth that kind of speaks to these adult fears considerations we talked about and then also this person added another thing I'm curious about more philosophically is what do they mean by now they have an extended phenomenological background for the model which suggests that the now would be extended like an extended now however in naturalist models which I assumed active inference or potentially artificial intelligence framework based models are a subset of time is often modeled and understood as point like or as a continuous variable with points I wonder what the implications are or what there might be if they mixed up these two interpretations of what the now is the physical and the psychological now so David what do you think about these questions from the participant they are great I'm gonna try to address some of it of course and again it's there are still a lot of things to discover and understand and one of the thing we like about starting with a mathematical model is that then it act as a guide now it's the theory that tells us you know where are the problem and where are the solution if you it's a general analogy sorry I'm gonna be I'm gonna I'm going a bit of track but it's useful and without considering ourselves with Einstein don't worry we don't have that ambition if you think about general relativity once the theory was the mathematical theory was derived basically people looked into it and started thinking there is a problem with this theory tells us that there should be black holes right and in fact basically the theory predicted things that nobody could observe of course before and very difficult to observe and that was the consequence of the mathematics and so that's the idea that we follow from a scientific standpoint is let's give us mathematics and see where it leads now one thing I'm curious about the first question yes we are working on more complex form alluded to that at the beginning number one this notion of transformation of as an investor of distance related to this literature is not to say that we appraise information only that way it was to show that type of relation holds in appraisal and to some extent or model makes sense of it now about the more complex form remember the space here the geometry is one thing but it's embedded into an active influence framework is a lot of complex computation involving higher beliefs and preference internal model of the body of other objects including internal model but we are doing now of beliefs about the beliefs of others and beliefs about the beliefs about the beliefs that other believe other have so basically level six of second order theory of mind and by actually imbuing artificial agent with this type of statistical processing and combine with the geometry then we can start modeling how you know inspired from simulation theory we can put ourselves in the shoes of other we can try to imagine what it would feel like to look at things to observe things to aim at things as they seem to do including understanding that they may have different beliefs and preference as us the same object may look cool to us and so when you get closer it's better but for them maybe that's something they hate and when you get closer they want to run away so then we have a way of extending the model very naturally with the same math with the same general principle to encompass theory of mind and social affective perspective taking if you wish and one interesting things that we find is that when you enable artificial agent with this type of process then you enable them also to be influenced by each other because you may be able to understand your but not care about it but if you do then the agents, the little robots or the virtual humans start showing joint attention behavior they start interacting, looking at the same thing looking at each other to kind of confirm that they understand what the other likes and they constantly put themselves in the shoes of the other and reciprocally and through this process of projective perspective taking informed by these kind of Bayesian-like models and the free energy minimization they start showing very complex behavior including we have reproduced in a non-verbal way the famous Sally and Anne task about theory of mind basically to puppet one of them they are together there is some object of interest in a basket one of them leave the other take the object of interest from the basket and put it in the box basically you ask children when the other come back where do you think she will go and if you understand false belief you will say well because she believes that the object is still in the basket she will go to the basket and so we have modeled a kind of simulation framework in a non-verbal way for that and we show agents actually exploit their understanding of false belief in others to approach disingenuously object they would like by being freed of being bothered by the other because now the other has false belief the same way we can kind of simulate some aspect of social anxiety because anytime you have an agent that believes that the other agent doesn't like it then playing the algorithm just running the algorithm the agent imagine itself approaching the agent the other agent that the agent suppose do not like him then he put itself because he practice social perspective taking in the shoes of this other agent and when he imagine what it is like for this other agents to see himself or itself get closer because of the same rule we have discussed he understand that would raise free energy in the other agent to account the other agent opinion that it attributes to the others therefore that raise its own free energy as a result it avoid it tend to avoid approaching the agent that it believe doesn't like it so you see you start having very sophisticated behavior holds through this basic processing so that's for the first question sorry it was long I hope it was not too confusing anytime you have to deal with theory of mind in particular second order theory of mind and talk about belief that another may have about the belief of another etc etc and so on it become very challenging for the mind so now regarding the other question about the now the spacious present so in a way it's a hard question to answer and we are still working on that in a way there is the physical time let's call it that way which is point like probably that is involved in the underlying processing whatever it is based on the brain or based on some computational algorithms so there is a timer and then there is the time as it is integrated, perceived and structured in consciousness and the idea here is that because we are doing active inference constantly anticipating the future to some extent even a very close future it can be at different scales so then we are trying to recover from that but we are not yet there the notion of this kind of extended present so that basically at a given time instant that's not the right at a given moment if you wish you not only have this instant and use representation of space in perspective but you have a process in which the anticipation drives the next instant in a manner that is integrated as a whole so it's very abstract what I'm saying we are also trying to understand how we can put time pressure within this free energy context and we have solutions for that and also how the agents can dive into the autobiographical memory under the same process we are also very interested now it's moving away from that but it connect to questions that we had before to the I would say pathological transformations in the sense that here we are talking about very adaptive behaviors very tuned to sensory matter contingencies and constraints and so that has this embodiment the fact that we tend to have this first person point of view refer to the representation that we have of our body in space that constraint somewhat reduced the number of possible projective transformations because only some of them will give you that type of point of view but if you relax those constraints then you can get very wild things that I find are very interesting to connect with the literature on psychotrop on mystical experience and arts and to give you an example one of the things that happens to the agents is that they have the possibility of basically being in the application of the projective transformation they can use for instance in the context in which they are in a desperate world in the sense that everything they can think about or do as an embodied agent will yield a very high level of free energy there is no good solution then one option is to release the constraint on the projection and when you do that in order to minimize free energy the agent will tend to basically put their point of view at infinity what we called the God point of view which is a bit of a bad choice of words and so basically that means that one way of minimizing free energy for those agents is to releasing those normal constraints is to escape towards infinity so that basically the world become very very far and then the structure of the representation of the world will start adding very odd symmetries that may remind us of mandalas or things like that so I'm shutting up now wow a lot of stuff there let's go Adam and then Lee and then Stephen hi can you hear me yep sounds good so one thing I've been struggling with in terms of time and consciousness is whether it's better to think of it as like a series of kind of discrete snapshots or more of a continuous flow I'm personally leaning more towards discrete snapshots and so with respect to projected geometric modeling I'm wondering whether the refresh rate might be like it's something like alpha frequencies like 10 times a second to primarily like posterior cortices getting into the mix of establishing that 40 perspective but then we're like and you mentioned like access consciousness where that might start and like the whole active inferential unfolding that that might be more like theta frequencies like 3 to 4 times a second and then like you might get something like north-off style like a spatiotemporal alignments based on like your interceptive states maybe more like delta frequencies and so you might have this like nested hierarchy of time scales with a core screening over events at various resolutions but at the bottom of it though it seems like the core of it is this establishing of a field with projected geometric modeling at the center I guess are these estimates I'm wondering discrete do you think or is it continuous that's really a very good question very sound and interesting hypothesis it's a whole program of research I've been for long an electrophysiologist I work for long on this stuff I don't anymore but I'm very open to pursuing collaborations on that first of all I should say I don't have an answer to that question I think it's in the multi-scale I think that it has the discrete of quote-unquote sharp if not discrete component which may be perfectly embedded in a more at least in the way it manifests itself a continuous flow but that is I don't think in and of itself doesn't pose a challenge a theoretical challenge it's just complex but we can think about how it could be done and we can also think about how it could be calibrated based on the type of biological you are talking about what is more of a theoretical challenge is this notion of spacious presence in which somewhere else that's the words that the phenomenologists like usur would use we pretend this notion in the now there is somewhat us in the center of that now and there is at the same time a potential towards the future and somehow there is simultaneously as a manifold an experience of the here and now and an experience of the just after and at the same time there is still this resonance of this retention to use in some words retention of of an immediate past so you can think okay there is an immediate past there is an actual present and there is a close future and then those things are somewhat bounded together but the difficulty is to try to think how actually you can find a mathematical expression of a representation of that in which they are actually part of the same whole and not just you know something you could argue well but then that means that they are ultimately separate process so we're still trying to work on that see if there is any way of thinking about that but we are not yet there cool, we'll do Stephen Lee Sasha yeah just building on the idea that the projective consciousness can be released when things are not certain enough to be able to know how to make sense of action in the world that sort of ties in with this way that indigenous approaches are shape-shifting and they're trying to attune to an environment which they aren't able to construct in the way that we tend to construct and place ourselves in the centre of so I think there is an interesting thing as well as there being the risk just be the increasing level of ambiguity and uncertainty in the world so that's interesting that you take that I'm just wondering if there's as well as the projective thing going from the self and maybe to infinity do you think there might be another process another type of geometry come in the other direction like it could be and this it sort of shifts towards the self-centred and there's an allocentric mechanism as well so that's such it's a great question and that's touched to another work that we are pursuing with Mathématician Daniel Benquin and Kenneth Gulliford who just told me unfortunately he could not make it today sorry about what sometimes people have called pre-reflective self-consciousness it's a bit along the line of this bizarre sense of the spacious prison but now it's about self-consciousness in the sense that that's what people have called sometimes the self-consciousness hypothesis not everybody agrees upon that that we are always consciousness is always conscious of itself that is it's not like a higher order theory where there is a consciousness for instance perceptual consciousness of something and then you reflect upon it and you know become aware of the fact that you are perceiving that the idea is that it's actually the manifold itself is such that you know there is this pre-reflective reflexive relationship that is fundamentally to it and so philosophers like Sartre I believe and others I've talked about to say that in English the seer seeing the seen seer if you wish the one we seeing seen at the same time as if there was some sort of duality in which whenever I have a first person perspective and I'm looking at the sky it feels like there is something in it that feels like the sky is looking back at me but as it turns out it's a natural property of the most you know bizarre, strange and interesting aspect of projective geometry which is not just you know it's a very complex and rich geometry and we're trying to work that out thanks to this mathematician around this concept of duality, reciprocity in which there is at once potentially we're still working out the math this is the fact that you take a perspective in implies that you know there are reciprocal perspective and that you know you look at something and these things look back at you in your I know it's odd it's bizarre but it's my my attempt to try to touch upon what you were talking about okay Lee, then Sasha, Adam thanks for all these cool questions yeah so I'm anticipating that I might be about to ask a stupid question but I'm going to go ahead anyway so it was stimulated by the fact that you were saying that you were kind of interested in you know like mystical experiences in relation to projective geometry and I was thinking about the meditative experience of non-duality which is I was wondering whether you have come across this and whether you see it in any way related to projective geometry perhaps yeah so I used to be I was at some point I don't know if you know that person is a student of Francisco Varela unfortunately passed away during my PhD that's why I moved to the US to work with Antonio D'Amazio to finish my PhD and of course with Varela who was very interested in let's say meditation in a broad sense of the term both for a scientific standpoint from a scientific standpoint as a possible method to study consciousness but also more at a personal and existential level I got more acquainted to that I practiced it even studied it a bit and I've been for instance involved in a very nice workshop in Israel on perspective on consciousness where this type of tradition multiple traditions we are meeting together and we had very interesting discussions and thoughts about that very nice group so can you try just for me and for the audience perhaps to specify the phenomenology the aspect of the experience you are describing the meditator, the specific meditation experience that you are describing this kind of unified nuance Sure, I can try so I think in some sense there was prior to a certain point there was an assumption that there's a separation between the observer and the observed and that starts to break down such that you kind of it becomes one space for want of a better Yeah, so I think that of course I'm not saying that's the answer I'm saying that perhaps that contributes to that type of state that connects to the what I was talking about that intrinsic to this reciprocal relations that are embedded into this geometry I wish the mathematician I work with was here there is something like it's not a trivial that's important a reciprocity between looking at or aiming at and being looked at simultaneously it's not like necessarily you lose track of the perspectives that you have but there is this more the richer experience that is implied by this type of geometry so in that sense at some level you can perfectly conceive within this geometry how perspective can be taken from anywhere and are all connected through projective transformation but even deeper perhaps through notions of dualities that I unfortunately could not expand upon because I would totally fail that's why I need to work with mathematicians that gives you this sort of unified experience of some sort of bizarre panopticon where you could be anywhere you could be looking from anywhere and from your particular perspective and being looked at that could be you that could be something else and that's you know a manifold of point of view that at the same time is grounded into a unique point of view that exists in full dimension so I know it sounds bizarre but that's just the maths that's what they are thank you we have a new visitor but Sasha go ahead then Adam yeah just thinking about this whole time perception question whether and I guess it's two separate things it's time discrete or continuous and the second question is time perception discrete or continuous and if it is possible for people to have different you know spatial understandings or perceptions then it's also possible for people to have different perceptions of time passing and different rate of how everything is happening and is that how do we integrate that or allow for that to happen that people perceive time perhaps very differently well that I don't think in and of itself is deeply difficult conceptual challenge at some level at least in the sense that it's not different to think about it as having systems that have internal model of beliefs and understanding that you know somebody else may have different beliefs in my robot in the example I was giving it was many beliefs about whether or not an agent likes or dislikes something but you know and all this let's put it that way let's be broader in the mechanics of the whole system there are a lot of parameters that concerns beliefs that concerns the calibration of projective frames that can be different from people to people so I would guess that it's not satisfactory I'm sorry but that could be the same for the way the flow of time is structured is construed within the consciousness you know at some point you are living with complex systems that have you know statistical models with a lot of parameters so it's perfectly logical that if you have parameters pertaining to time consciousness they could be different from one person to the other you know at that level of abstraction I don't think the problem is an essential challenge in the sense that here there is a big paradox that I cannot resolve in a consistent manner with mathematical thinking but of course there are much deeper questions than this whole what we were talking about before this whole spacious present this kind of extended experience of experience of the present of the now is more difficult to grasp to me yet let's go last comments here from Adam and then we'll welcome Ken hello can you hear me so in talking about this sort of reciprocity or mutuality of like when you look and then looking back at you it started to make me wonder about the extent to which projective geometric modeling might be intersubjectively bootstrapped and that like it both enables theory of mind capacities in terms of giving you this flexible transformation between perspectives which would be like at the heart of like mentalizing a theory of mind but also to engage in that I'm wondering if that helps to induce the capacity which then feeds back and makes you better perspective which then kind of bootstrapping process and now I'm wondering in terms of like evolution and development like when is this starting to arise like you could in theory think of this happening like non intersubjectively in a way where it's like I don't know like a predator sort of like hunting maybe if I go here and doing sort of like where do I need to be to get like if the organism has like a model of itself with some degree of temporal depth and it's trying to do something a little bit more labyrinth trying to like forge for information unless okay I go here I'll have a different perspective and this is like influencing the policy selection but in humans it's probably really heavily intersubjectively bootstrapped I know so some thoughts I'm not going to make a long comments to that take too much space but I find everything you say that I agree with you I think actually you right probably so they are very interesting stuff in etiology and you know animal behavior or even in recent relatively recent literature that you know suggests the role of projective representations into the way any more try to lure others into you know either being a mate or being a prey cool welcome Ken you can say hi yep sorry I'm so late all good do you have anything to add or just feel free to raise your hand anytime okay we'll do well there's a few more slides this was really a great set of questions by the participant so that's really fun topics here's some quotes that I put up there and the part I wanted to kind of start off with is the second half of the second quote on here which is just a system running the PCM would more closely approximate the actual complexity organization and processing power of human brains and this could very well on the author's view literally be an artificial consciousness so maybe getting into sort of questions about what is it actually entailed to have the awareness what if you make a system that has action and perception and it does a projective geometry transformation and it you know it's doing it because you're the one who made the robot so I was just curious to hear about from the robotics side or from the mathematical side or from the neuro-phenomenological what else is happening in humans is it evolutionary is it linguistic etc so I would like Ken to speak because he has a lot to say but since you're asking about robot in that question is very relevant of course I will answer as a scientist I'm hiding behind the method the scientific method not to avoid the question but to tell you first that I don't have an answer to that fundamental question what I'm trying to do and that's I think the only thing science can do at least as we know it and you can take the model of physics as a prominent example of that is to try to understand not necessarily why but how we do think we do things and that applies to consciousness to the phenomenology of consciousness so what we are modeling here is that phenomenology and how it is involved in other process and so at the end the only thing I could say about my robot in the way they are implemented at this point you know there is a lot of work ahead is that at least their behavior will be governed by principles that are central to the way consciousness process information let's put it that way so it doesn't imply at all that those robots are actually conscious in the sense we would like to but at least they can demonstrate how our consciousness and the principles that seems to shape our consciousness play a role in overall the economy of our behavior etc etc it doesn't give us an answer to why you know this is conscious in the sense maybe actually with a certain level of complexity it's sufficient and maybe Ken should speak now and rebound on that well I was going to say this passage if memory serves of course this is a general sort of worry that anybody who works on consciousness thinks about it has especially if you are doing anything with simulations as David was so we started to think well of course we don't think that the simulations that David has built are actually contained consciousness actually contain conscious agents so then you wonder all right so it's not merely processing information in the way in a way that simulates what's specified in the model I guess you could say it could be other things as well like so in a sense we think okay this is a necessary condition for consciousness that you process information in a way the PCM outlines sufficiency acquire a certain kind of scale and complexity that we we're not certain about yet so we shouldn't just assume that any kind of solution that follows this model is just so facto a conscious being that's kind of claiming too much what else is involved we don't know but it seems like we maintain this is necessary for consciousness so a conscious being will have this structure process information in this way and for all we know it could be sort of things that we don't we're not going to be able to reduce to principles so easily that we'd have to discover empirically that's the reason we say here an a posteriori identity theory that's the idea that say what consciousness is a priori you have to find out what it is empirically and at some point you might find yourself in a situation where you could say okay these suppose we build a lot of artificial system that do this these exhibit all the behaviors and so on that make it seem like a conscious being they suppose they operate on a certain scale and a certain complexity at some point we might be able to say yeah we think right here's the sweet spot why exactly we don't know maybe we will discover that maybe not but on the one hand it seems like you don't want to be shy of saying that from what we can tell these are necessary conditions and on the other you don't want to be to you don't want to have a view of the sufficient conditions for consciousness that make it quite too easy you know awesome yeah thanks for that response Steven just a really quick comment and then we'll move on just quickly I was just wondering when you think about projective consciousness is it like this is really summing up a particular type of consciousness which we are taking these projective perspectival kind of ways of engaging the world and there's a potential I wonder if there's a potential for that to seem like it's more like how consciousness is because we culturally have taken that on to be how we think of the world but this might be like one way of view from a more sort of higher perspective so shall I take this one what's that what's that I think you should speak and because I have spoken already a lot and you have a lot to say what's your name so I mean our hypothesis is that this is not just of how we've been inculturated or let's see if I got your question right our view is that this is at least a necessary feature of all consciousness that's the claim so now you could imagine you could actually wed this kind of view to other theories of consciousness you could wed it to panpsychism for that matter you could just say oh no actually consciousness is something else it's whatever some fundamental property of nature but then if you had that sort of view this kind of work would still be of interest because it might sketch out how conscious beings like us or probably animals how they experience the world and how they process information and of course when we make it so it's compatible with that compatible with really consciousness that's cast at a kind of metaphysical level I guess our view would be something like well suppose this type of model allows us to explain everything that we think is pertinent to consciousness so it allows us to account for whatever behavioral and cognitive functions consciousness carries out and for the phenomenology of consciousness if it allowed you to account for those things then and you suppose you could identify the type of structure this is and the sort of scale and complexity it needs to be organized on then to our way of thinking at least as it comes to the metaphysical issue we have a tendency to think okay that's just what consciousness is we discovered this empirically we discovered it opposite to a riori there's no need to postulate any further metaphysics because any metaphysical account of consciousness well it actually has the same structure it's something you didn't figure out a priori really you just thought okay well here's an account that would make sense of the fact that take panpsychism why is anything consciousness well you know conscious well if everything is conscious then it's no mystery that we are so that's that's an argument you can make well you know as you probably know there are all these different positions in the philosophy of mind about the nature of consciousness and the distribution of consciousness in my view they're all kind of epistemically on a par none of them are really like you can't show any of them a priori you can't demonstrate so it seems like it's wide open to you on the theoretical level to be this kind of identity theorist just and you reach you know we have this structure have this level that your conscious being that's what consciousness is we figure it out empirically and we don't need to sort of postulate anything else that said it's totally compatible with other views so you could hold this is much more derived and primitive like sensing or something is basic form of consciousness or something like that we do think though that so here's one thing that's not exactly a priori but let's say it's a strong intuition that a lot of people have if you're conscious of anything that's going to involve the appearance of something to so it has this kind of appearance of structure and that's a kind of spatial spatial I mean quasi spatial structure a proto geometrical structure and that takes you already kind of in the direction of thinking of consciousness as like a space and like a projective space in fact that's not you can't demonstrate that a priori but it is a strong intuition that people have so I'd like to say and I think I say in this paper don't I you know follow your phenomenological muse build your models and see where they take you because we can't really determine those things a priori or just by banging on the table with our favorite Thanks I see Adam and Lee have their hands raised but just really quick would especially David and Ken would either of you be okay with staying for like a few minutes past the hour so we can just continue talking for minutes okay I have a call in about 40 minutes so I'm fine oh beautiful so we'll just anyone can depart whenever they need to and we won't have to come to a speedy landing there but we'll go Adam and then Lee oh hello I guess I'd like to hear a little bit more about the intuition for why does that existing simulations might not be conscious and in terms of what it would actually take to get a sufficiency this is something I've been struggling with also the for me like I've been like mostly coming at it from like the neural direction thinking of like okay what in terms of neural systems might underlie these different capacities and then like what how would that be implemented I become a little bit more confused when I start to think of alternative implementation so like something like um projected geometric modeling I'm wondering if that's primarily like a posterior meal courted posterior meal courtesies and that's like giving you this sort of flexible spatial modeling but then this is like coupled to body maps of like lateral parietal lobes and that this is part of like one basically synchronous complex which is mutually constraining each other as you're filling in this sort of joint estimate of the world and your lived body in relation to it so basically exactly what you propose and so for me it's like when I'm talking about a biological system there's a certain grain at which this calculation is being established through like a metastable synchronous complex or I call it like a non-equilibrium city state distribution or I call it a self-organizing harmonic mode but if we're talking about something like a simulation on a Turing machine that might be doing the geometric modeling by very different means is that an issue is that not an issue for you just curious to know more thoughts on a boundary problems thanks maybe I can give my thought about that and then can. Number one about the functional neuroanatomy and the neuro dynamical mechanism that may implement the type of process we are talking about. I think I agree with you. In the paper we published in 2017 in the journal of theoretical biology we have at the end a little figure that kind of summarizes our gross functional neuroanatomical hypothesis and indeed it seems that the posterior you know in particular the parietal lobe and to some extent the posterior cortex but the posterior cortex maybe more is a bridge towards interaction with other system but the parietal lobe and then we could enter into the detail is certain is probably where this sort of project as we call it in the paper projective geometrical engine operate and I think the way you attempt to make sense of the difference between something like the model being processed the way you describe it into a brain versus into a Turing machine basically a computer is perfectly fair and right and maybe indeed when you do it in the Turing-like way like when I do my simulations with my virtual humans and robots you know it doesn't entails the type of experiential aspect that we are trying to figure out of which we are trying to figure out the course basically the sufficient condition and perhaps indeed it's a very different context of processing the way you describe it and that is actually key perhaps not you know it would need more elaboration of course but to make that process conscious so for me it's not a problem it's a different type of implementation we can use the Turing-like to do simulations from which we can learn a lot but probably indeed there's something about you know the embodied processing with this type of whatever the brain is doing and you know I'm a neuroscientist initially and I can we all know that actually we know much less about the brain you know we pretend to and they are probably great and important deep fundamental research to understand and until we have understood the brain in a much deeper way well we won't be able to conclude on that question Ken? Yeah I can just add that you know I mean suppose it sounds a little odd you know from the kind of computer science perspective to think that say performing all these computations in a massively parallel way so that you know I mean you can imagine you know a tinkered computer doing the computations involved in taking like thousands of years or something and you think well okay that's it's not merely running this program right but if you say okay it has to do with massively parallel integrated computation of these functions that of course sounds a bit like like strange like why would simply being able to do it really fast or do it in this parallel way make it make a difference but you know if you're serious theory or identity then you think well okay we'll just find out right it does it does seem really strange though to think that a Turing machine made out of beer cans and string is John Searle carrying out these computations would be consciousness so that's why I think it's a good idea to leave a little bit of space here so that's what we try to do maybe let me just add a little historical note it's not about Turing but it's about von Neumann we need a seminal paper in the Bell Laboratory I believe and he derived he explained the foundation of his architecture that is now basically an implementation of a Turing machine into the computers that we use currently actually and then you know it becomes like oh von Neumann machine versus the wet brand but in fact von Neumann in this paper start from models of artificial neurons the basic maculoc and pit neurons to try to figure out for basic computations such as you know arithmetic operation could be abstracted and then he proposes a serial model of computing emphasizing that at the time because they had to use those vacuum tubes to make those the equivalent of semiconductor you know it would be too too big to build a parallel machine it would just not be physically practical you would need like a giant building to put enough tubes in there and so the solution of doing it in a binary manner and in a kind of sequential manner it was just a practical solution but he's thinking was broader so you know maybe there are also deeper aspect of computations we don't understand yet and that could help us also to make sense of the experiential dimension of computations and if I can add one thing to that it's also true that simply massive complexity hides I mean it has sort of I don't take this too seriously but you all know the old sort of Hegelian idea that quantity eventually yields to quality so you have quantitative changes and you have enough of them with transitions to quality well there is one way of thinking about that not taking the metaphysics of dialectics seriously but looking at it epistemically where very complicated things that end up being in fact this is kind of the way we are oh somebody brought this up okay well alright so I'm speaking to nobody's gonna be freaked out Hegel anyway the idea is that particular notion of going from quantity to quality so just think of massive complexity you take a massive complex system as a whole it can behave in ways that you have to understand sort of qualitatively I mean even things like chess playing computer programs to take the classic example approach your chess playing computer if it's a good program from the intentional stance from Dennett's term you regarded as trying to attack the stream and so on you don't think of the low level complications going on in it to implement that so that's all just to say massive complexity with the massively parallel it's not totally crazy to think that that could yield something that has this kind of form and this kind of macro structure if you like that we experience qualitatively that consciousness has it's not completely crazy to say okay yes serial computation couldn't do it but massively parallel computation could I've been thanks for that that makes a lot of sense to me and I guess in my own thinking I've been struggling with to what extent I'm dealing with practical versus in principle issues with respect to boundaries and what it would take to realize consciousness the massive parallelism is practically necessary but is it in principle necessary and there does seem to be something strange like church-turning thesis in theory you can do it with the beer cans and rope and let it just play on a really long time scale and eventually you calculate the whole darn thing but you're just like I don't know about that but why so it would be like practically there might be a sense you need consciousness to both inform and be informed by action perception cycles but that's still a practical not necessarily in principle but then I guess I've been wondering if it's for going to in principle would be like is the computational objects real or not so like if like a simulation versus an emulation or things like this and with the recurrent dynamics sort of like metastable synchronous complexes model I'm working on even then there's like an issue of like is this a real thing or not and like they're dealing with this issue with respect to Donnie equilibrium study states and the free energy principle which is like these things are defined as these like harmonic modes but then their boundaries are like they seem to be model or dependent like what or like in a neural system like do you ever get like exact synchrony how much is enough to count what's the time scale at which the thing actually has its existence as a particular attracting state and relative to what and so this is where I start to this difference between practical and in principle issues of implementation and what would actually and then what would let you say okay this thing is conscious I'm still really confused about it may I interject may I comment on that yeah I'm going to go alright I don't seem to have a opposition so I started developing an hypothesis a while ago that I published in a journal called advances in neuroscience that actually died I don't think it still exists and it was kind of based on a reflection on my own work and the work of other in neuroscience and psychology not doubly working with brain damage showing that you know a lot of what we believe as being critical for cognition in many aspects of you know cognition and affective processing in the brain like everybody believes now that the insular cortex is critical for craving and many aspects of the sense of well I worked with brain damage patient that basically had no insular on both sides like 40% of their brain and nevertheless you know except for some deep entero-grade amnesia that could be explained by their whatever they were normal so I came to actually think that the brain is much more complex than we think that they are the way we observe it with the method we have no imaging or electrophysiology is very low level it's not different than putting thermometers on a microprocessor and trying to infer what type of operating system is running and so I started thinking so maybe there are you know layers of abstractions that we haven't understood in the way the brain process information and one analogy that came to my mind I was very interested in computer science at the time is the process of virtualization in computer science that is basically a way I mean it's a simplistic way of explaining it a way of creating abstraction layers so that actually at the user level of course we are talking about computer involving user but at the process level all the details of the hardware are hidden you know when to take a very trivial example that is not a good analogy when you use your desktop computer you don't see the the machine language you are not using word you don't need to program assembly language to use your Microsoft word interface so virtualization is actually a process through which software or process can be abstracted hide the complexity of the hardware and also become much more resilient in the sense that then you can kind of move things around across servers across you know different things so it's not an answer to the question but I still have this intuition and believe that perhaps there is something analogous that is bridging the gap to some extent between you know what we know and observe of the brain and subjective experience and more you know higher order level process that is analogous to virtualization that basically the phenomenal experience of consciousness is very useful for active inference etc but relies on neural mechanism that precisely are designed to hide the small and complex details of the hardware and therefore makes it more usable more practical and of course at the same time it takes longer because you need to run many cycles to do that for instance if you want to simulate a certain type of microprocessor over another architecture of microprocessor etc etc so I think there is something there but trying to understand is how the brain may create a virtual space and through virtualization process that then may help bridge the gap between the different domain you are talking about even though it's not sufficient yet in my mind it doesn't fully address the problem cool just a couple notes to tie together a little bit of the discussion so we've heard about a few classic philosophy of science dichotomies like necessary and sufficient practical and principled from philosophy and then also like real versus unreal seems like to some degree brains would have to because they're so complex they would have to create these abstraction layers just to work as control systems just to work as integrated systems they would they'd have to like you'd have to self-course screen just to actually have purchase to do like hierarchical control and then what you're doing this then you might get potential some like flexible multiple realizability from that potentially or non-flexibility too depending on like how you did it but you summarized Adam again very good summary of a better presentation of my own thing well you've influenced me very heavily so I don't think that's a coincidence yeah one just to bring a few of these threads together this is all really interesting stuff so one is that in complexity we say more is different and that's related to this qualitative to quantitative more quantitative changes can lead to qualitative differences through phase transitions self-organization just qualitatively different behavior strong or weak emergence from just quantitatively more and then that's always going to be a relational question how many grains of sand are in a heap that's the whole point is that's not pretending to be an objective question it's about pretending to be a non-relational question even though it is about the world and these ideas about which levels of the system are available for introspection versus is related not just to attention schema theory and attentional salience from the FEP and active inference but also a lot of things in Girdel Escherbach and so there is a Hofstetter citation and a lot of these ideas have a lot of analogy there which just shows how in some senses many of the same questions are returned to even 40 years later so we have to balance out the big computing hype from the really quintessential philosophical questions like necessary, sufficient, practical questions, principle, real versus unreal her spectival and taking that seriously, the relational perspective integrating topology and geometry so it's just really a great discussion and that the way the paper and actually the sequence of papers is written is very clear interfaces and highly concerned focused with different metaphysics but just saying it's neither simply necessary nor sufficient to have this projected mapping of your condition X and Y for example the vanishing privilege point and all these things we were talking about it just goes downstream from there so it's an interesting approach and as we're hearing from this discussion there are a lot of questions that come together in thinking about these topics so thanks for bringing all these perspectives to the table so this cool stuff maybe Ken can add on your reference that you had before how Hofstadter works and get a lecture because actually Ken has worked a lot on these questions that actually contributed to the strange look book of the Hofstadter by editing a version of it and he knows the Hofstadter so maybe he can add things to say cool, this is the section where one of the sections where Hofstadter was cited but I'd love to hear about this Ken it'd be great, thanks well it's just a little bit of I guess history about this actually I met David through he and Antonio DiMazio were contributing a paper to a book I was editing and Douglas Hofstadter also had a paper in that book and the papers were supposed to be like I don't know 30 pages maybe and Hofstadter turned in 100 pages and it's kind of typical way so I said well look we have to edit this and he's like okay but be very careful and so I went through this long process of editing down his contribution I cut it in half and that eventually became I Am a Strange Loop which is cited here but he was ultimately okay with my editing job and in particular what I thought was really ended up being really nice about I Am a Strange Loop the way in which Hofstadter is able describe basically the causal efficacy of implementing mathematical forms at least in philosophy think of mathematics as a totally abstract platonic way doesn't have any causal powers and so on and then of course there's the old much there's the Aristotelian tradition which says no no the forms are actually in the matter they're implemented in the matter but that position is not particularly you can find it but it's kind of like it's not the dominant sort of thumbnail sketch of this issue in the minds of most philosophers but I found Hofstadter's book to be really good on that implementing certain computational structure certain ultimately mathematical forms has real causal power it does change the system that implements it and it's perfectly legitimate to think of it as a causally real thing one example from Hofstadter's writing that reminded me of today was when we were talking about whether it would simply be enough to have a data processing model that involved a projective geometrical component would that be necessary or sufficient for consciousness and your very articulated answer was about how actually we know that these are some sort of features that we want to be modeling but they're neither at the necessary nor sufficient tier they're things that we're seeing as sort of being spun off from some underlying process that there's still a lot of mystery around and so that's like the phonographs with the low resolution phonographs in Goethe-lescher Bach it's like you can make a very limited thing that does one thing without error or without some of the fancy attributes but then to extrapolate from this less is different more is different case it's like well it's not a phonograph okay it doesn't play records yep well it's that kind of a phonograph that's what it's like to be that kind of a phonograph that's what it's like to be a light switch okay well that's not really the complexity of system that we're talking about here which is the embedded in culture human and then that level of complexity the full featured phonograph opens itself because of its strange loop nature to this whole new universe of recursion and of loops and of self-reflection projection and then the phase transition or more is different question is like how do we get from things that don't have that strange loop feature to things that do but there's nothing that is sort of a strange loop or is not a strange loop it's like either it loops back with sufficient complexity to be strange in the world or it's just not strange enough to be as interesting as these open-ended infinite generative systems that we seem to be participating in Adam and I'm curious about how this interfaces with free energy because I don't think there's been a half-set or free energy crossover quite like this Adam Hi, I'm thinking again about what David was mentioning with this sort of like reciprocity or mutuality of perspective of like seeing and then like looking at something and then this and being looked at and the ways in which some of these things might be intersubjectively strapped and I guess I'm wondering whether that puts like a strange loop like right at the heart of this model and I'm also wondering whether one of the things that might potentially get you is a recursion and a powerful kind of reflexivity which would then you can then start talking about things like conscious access and all sorts of metacognition kind of like growing out of that potentially yeah some thoughts Can you need to answer that question in reference to self-reference and non-wealth from that set theory and whole project you are the one to address that Okay, well so actually that volume that David contributed to and it was a book called self-representational approaches to consciousness and it's most of it was written by analytic philosophers and some a couple of phenomenologists but that was the theme basically how people who think that I long have that consciousness involves this kind of reflexivity kind of self-awareness that it's not just introspection it's a more basic notion of reflexivity than being able to attend to yourself or reflect on yourself and that book was about that but in that book I kind of looked at ways of formalizing this using so the Hofstetter I've asked Hofstetter to contribute to it because we thought that's right in line with Gertler-Lescher Bach and the notion of kind of using the notion of metamathematics representation in self-reference in the metamathematical context as sort of one model for the kind of self-awareness that's involved in consciousness that's the kind of quintessential strange loop for him with respect to consciousness but I looked at models using hyper sets like sets that contain themselves as members and now recently though actually working on that is kind of at the origin of the whole project the whole PCM project because I moved back to Iowa City in like 2007 I think that's right and David was there still in the situated from the neuroscience program but he was still working at the neuroscience program at Iowa and we kind of formed a little research group just based on this and this eventually we kind of discussed geometrical models a little bit but David really had the breakthrough with regard to the game of projective geometry as the way forward somewhere around 2010 something like that but we long wondered how can you really connect these two threads the geometrical part the projective geometry and this notion of well to use Hofstetter's strange loopiness of consciousness and we're actually still working on this so we think we're not certain but we think we thanks to Daniel because Daniel Benka we think we found a way to connect these two notions finally more clearly David do you want to say a little bit about that about how the PCM can sort of make sense of this notion of strange loopiness strange loopiness so that was really a driving force in the end in trying to select a model that the model would incorporate intrinsically properties like that when I was trying to because I always had been convinced that there was something geometrical about intuition so we needed to move from kind of abstract graph theory and even with this kind of non well funded sort of set theory that allows you these loops etc to be more phenomenal more experiential, more concrete so I was looking into kind of possible geometries that had weird features beyond this perspectival intuition from which we introduced that and this projective geometry geometry under some boundary condition but actually it's intrinsic to it as a natural loop if you wish what I'm going to say is going to sound very awkward and it's a kind of qualitative description of things that need to be expressed and formulated rigorously so sorry for the mathematician that may listen if I'm doing it the right way but basically when you look at a point at infinity in front of you okay so the vanishing point just in front of you that's the same point that's the back of the point that is actually the untypedal point in the back untypedal point so two opposite points at infinity in that projective space are topologically glued so that in some sense actually if you were kind of a virtual little creature moving into the space of consciousness when you reach the point at infinity you find yourself on the other side of the space and if you think about it it's just as if actually the projective space you know this when you kind of start messing with or considering the plane at infinity and its topological properties basically was kind of a sort of mirror in the sense that basically when I look in front that's not the case in our visual experience obviously but that is in the structure of the geometry I look quote in quote of myself looking in front from the back in a kind of reverse way that's the property that is actually intrinsic to projective spaces but that we don't normally see okay cool stuff here thanks for sharing all this history too so just two thoughts one is this virtualization idea and in Gertel Asherbach it's prelude ant fugue Aunt Hilary the ant colony and is it a virtual machine well there's yes and no in one sense lower level details are shielded to the more semantically operative top levels so that's virtualization for sure concurrent programming as well but is it a virtual machine in the sense that it's a dot vm that I can run on a different operating system no that's not even the ant colony OS that's not a strategy that transfers so the virtual machine metaphor captures one aspect perfectly which is that there's shielding of lower levels of the quote assembly code but also there's another feature of virtual machines that they've come to mean in the modern computing environment which is like reproducible computing and that's definitely not the case that's because we have this embodied inactive ecological element so that's kind of a cool extension of a Hofstetter idea and then this idea of infinity being on the other side it's sort of like you would if your light rays were being bent around the sphere or whatever shape we're on it would be seeing around so that's straight out it like a laser so that's sort of an attentional laser reflected by the higher concentration of photoreceptors the increased sensitivity to observations there but that being the point of attention and even meditative practices often focusing on attentional points and visual rays like that but it does if you think about the ray of light going all the way around it would be the trajectory you would look at to see yourself from the back so very interesting there and that's related to our conception of the space of the earth whereas on a flat plane it would just go to infinity and then pop through in like a piercing way because the space doesn't bend back on itself Adam? Sorry So I'm thinking that you might be providing an additional account of some of the phenomena that Graziano tries to describe with attention schema theory in terms of certain intuitions about the nature of consciousness and self-hood and a kind of awareness is like a soul-like ethereal non-material intuition and so for Graziano, for him consciousness is this low-dimensional sketch necessarily reduced dimensionality for the sake of control to have a model that's more handleable that throws away details and gives you purchase this low dimensionality to it is like the thing that makes you that if this is your awareness and your sense of consciousness and self-coming from this this is part of certain intuitions about immaterial souls and things like this but if your perspective inherently involves this kind of strange loop this difficult conceivability where it's like you are both the constant context for everything the reference point but you're the one thing that's like invisible and in this kind of paradoxical way I'm wondering if even I'm wondering if that potentially contributes to similar aspects of phenomenology where like certain aspects of kind of like mystical intuitions about the self and consciousness some thoughts cool question and we're taking it nice and interesting areas David why don't you go ahead Ken could say a lot of things sorry somebody else wanted to talk you or Ken go ahead I'm going to just say something but Ken will be able to better elaborate there was a if a mystique or I would say a zoterist named Rudolf Steiner right Ken who actually believed that projective geometry because of this sort of reciprocity of perspective when I look at the sky the sky look back at myself was a connection to God we don't endorse that necessarily but there is this story of mystical and esoteric traditions that invokes projective geometry in the past in that kind of context right Ken yeah so there's there's the Steiner tradition which one of his students I believe she was one of his students Olive Wichier she wrote a really good production to projective geometry which I highly recommend looking at it's got wonderful illustrations and they seem to have been on to the idea that projective geometry was crucial to consciousness and mind but you know they also had a lot of strange other notions that made them seem like cranks and then there was another interest in just higher dimensional geometry that came from people like PD Ospenski if you've heard of this guy he he was he wrote the fourth dimension and his view is that there must be part of ourselves that we don't have access to in the way in which we have access to three-dimensional objects which kind of and he thought that would be in the fourth dimension it's not clear that he thought that it was a projective geometrical structure but still it's sort of in the ballpark just one note there is that Ospenski who studied under Ger Jeff deep esoteric threads and of course going much deeper time alternative ideas about number geometry color and of course philosophy those are in many ways what define the less religious at least in the exoteric sense of these wisdom traditions so it's very interesting to see the geometry spiritual connection return after who knows maybe never even a separation but at the very least in most people's mind they think of it as shapes on a plane not about the conceptions of the lower level structure of reality or the higher level structures of reality like the shape of the universe shape of things shape of ourselves so very nice threads and David's a big fan of what's his name Matilla Geeker right poet mathematician and Romanian Prince who wrote any book about the connection between mathematic aesthetics and natural form yes there are a lot of connections that were very inspiring to me my father used to show me his book when I was a kid and my grandfather was an art historian and he was also searching you know interpretation of art forms through mathematics so there is a long family tradition that led me here so I'm sorry but I will have to leave to another call it was very very great I can leave you can continue to discuss I was very happy so we'll get in touch and stay in touch and I'll see you bye bye guys see you all move us to the end but really David and Ken thanks so much it's like always just great to have the authors on here to be hearing all this extra stuff so the other slides just to you know visualize them people could screen CAPM is just a few more quotes to get to that Hegelian dialectic which we actually got to talk about we could have had a few more just general questions especially where does active inference fit into this whole picture and that's in the cybernetics and how it's carried out but active inference live streaming what not that's always what we want to think about and then we just had figure one figure two figure three and figure four so those are just things that we could have talked about maybe in the future or in 2021 we could have the moon illusion or a deep dive on one dimension or another because this was such a fascinating conversation so if Adam or Alex or Ken maybe each of you want to have just a last comment that would be great so let's Alex if you want to go first then Adam and then Ken or pass no worries if not okay okay I'll pass by now thank you Alex Adam any last thoughts how is it going to change what you're going to work on or think about or just um too many thoughts but one of the things I'm I'm doing a lot of thinking about like what neural systems might potentially contribute to different aspects of visual spatial awareness and geometric modeling and one issue I'm wondering about the extent to which um at least in its visual nature it might always the conscious awareness might always be um more of a 2D projection and that things like three dimensionality and things like additional depth and additional dimensions might involve um the ability to um transition between different points of view in a coherent way and a way that would imply a 3D world but where at any given moment the visual aspect is actually 2D potentially without clear depth these are some things I'm wondering about and I think it's not incompatible with projected geometric modeling it would just be whether the engine how distributed it is and the granularity over which it's being achieved and so this coupling between a 2D projection screen and a body map adjusting what you ought to see or something like the posterior singulate like saying here's where your eyes and your head is giving, putting your perspective in there and then it seems like 40 geometry could be perfectly compatible with this but I'm trying to figure out how that works there's a lot of rambling but thank you. No great, thanks for your participation and this figure too points out some of what you're talking about that's a projection onto the 2D there's an ambiguous projection on the left because of the perspective of being ambiguous and then free energy is invoked as like a symmetry break that gives a perspective on which side of the box one is actually on given a prior that about how certain angles must be reflective of 2D reputations so yeah Ken, any thoughts on that and or just a last comments Well just thank you for organizing this I'm sorry I couldn't make it for the whole thing and it's really been a wonderful conversation and I'll just say the next thing we're working on Daniel and David and I is a paper trying to relate the projective geometrical model close to show how it relates to notions like strange strengths and pre-reflective self-consciousness it's kind of more basic type of self-consciousness than introspection or attention and to show that you can make sense of that in terms of the projective geometrical structure that's the next thing on the agenda for us cool sounds like a fun research direction so with all of this being said everyone thanks a lot for participating live and if you're watching for watching it too we will have the follow-up form link is in the calendar invite if you want to just provide us some thoughts any feedback suggestions and questions are always welcome other than that just stay in communication really appreciate this great conversation and some awesome stuff so thanks again everyone for participating and watching and we will talk soon so see you later thank you