Textbook Group meeting for Parr, Pezzulo, Friston 2022 .

all right welcome everyone it's October 24th 23 and we're in our first discussion of chapter three for cohort 5 so before we go to any specific question does anyone just want to give any first thoughts or Reflections on three just where it is in the textbook or um what's stuck with them or what they remember from the journey Oli want to give any first thoughts or anyone else yeah sure uh well chapter 3 as opposed to chapter 2's um kind of bottomup approach or the low road uh takes the more daunting or uh let's put it this way the more adventurous route of uh the high road uh which is the one that's been increasingly uh developed in uh the recent years uh so it begins with the notion of what it is it or let me put it this way if we suppose that quote unquote things exist in an environment or in the World At Large what characteristics uh should they possess to consider them as ontological things in other words uh what character characteristics should they possess if we want to uh kind of re-evaluate or remeasure them uh at a specific time scale uh and uh consider them as cons consistently the same thing so the idea is if we put a kind of statistical boundary between the internal states of that object or the thing and its uh external States or its environment then it would be possible to uh to underwrite the Dynamics of um of the of the internal states of the system uh at face of the changing environment given that Mark of blanket or this or the statistical boundary so in other words the internal states of the system um does not change directly um in a response to the changing environment but the relationship between those two states are kind of they kind of vicariously uh influence uh the external States vicariously influence the uh internal state so they don't have direct access but only through the blanket States or Mark of blanket so that's basically the idea for the a self-organized object um and uh it's it's uh one of the main main properties for any living thing or uh rather an idea for uh how life Works um I mean by life here I mean in a very broad term and not just necessarily biological life but it also encompasses any kind of artificial life as well such as xenobots is developed by Michael Levan's lab or any other uh organoids or um similar uh entities that's been increasingly uh appearing in uh many various disciplines and research areas thank you that's a great summary Kristoff then anyone else who raise their hands um I just wanted to ask because the the concept of um Mar Blanca is sort of a separation between the internal and and external environment it always makes me think is that is isn't it really kind of talking about uh temporal SE separation that um ultimately what what really says that we have states that we can identify that are not directly influenced by uh by the environment but but ultimately the system is it is connected so um it's it's it's this this separation like you know eventually breaks down right the signal the signal gets through the barrier and finally influences the internal States there's a lot of pieces there first to last piece yes the signals do get through they do influence so there's causal propagation that can reify the blanket so blankets can exist as interface they do exist as the interface they also only exist over a certain spatial or temporal time scale and then you in the beginning described like time in a maravian sense and that's like a more common usage of the maravian property is through time like the past only influences the future through the present so then the current moment is a markup blanket in time so you could make a markup blanket around any Baye graph you write the Baye graph out you pick the node as internal that's the system of Interest internal states there is a base gra um there is a mark on blanket so that could be um through time you could have one variable through time and talk about maravian property through time it could be spatially embedded and then it…