Textbook Group meeting for Parr, Pezzulo, Friston 2022 .

all right welcome back everyone we're in cohort 7 chapter so for some of us this will be our first time on seven others not not so first um but where would anyone like to begin with a specific quote or question about seven or how did anybody take seven or where do they see seven in the book maybe I'll say uh I I listened to the uh video of the last session the evening session where Andrew talked about chapter 7 and so what interests me uh and I'm still struggling how to code things like free energy principle and how to um you know how all the math Works Etc but um and he did he did help explaining that but what I got from these sections was that they progressively um are looking at modeling different parts of mental life um and showing that in each case you know the this apparatus um uh can be used to model that and so in my case uh like so basically like perceptual processing like deducing an intended signal uh the teamas with the rat like decision- making planning uh epistemic versus pragmatic and then information seeking with the isaad uh seeking position where you can get it kind of yielding a street light effect and then learning a novelty so this synthetic worm exploring its environment changing the generative model and then finally like hierarchical or deep inference words and versus sentences putting that together together on separate time scales so so what it seems like is that um you end up you you know you you see oh I could use active inference for modeling this um and then um you do things of you know whatever sophistication you can and so I'm interested maybe in the opposite to say okay I'd like to have develop a cognitive model I think that I'm going to have like a continuous part and a discrete part like a procedure ual continuous mind and a discreet declarative mind and then I'm going to maybe Wonder like how they discover critical points navigate them um you know have a language of critical points so so I guess I'm just still learning about U the um active inference but I'm trying to pick up on that connection between like having a mental process or experience in life you're trying to model and then thinking okay where does predict addiction error come in I guess that's probably what it's about that's where I'm at right now thank you awesome summary anyone else want to give some thoughts on seven or like a a section or a quote or a figure or equation that they want to jump to um I mean figure 7.2 with the the plots about uh The Listener are a little confusing I guess um I'm not sure what I'm seeing in the First Column of panels there maybe someone with some more insight could kind of walk us through what these panels are showing yeah it's a great point because this chapter is the discret setting but then it's like wait a minute why are there continuous traces here with a discretized model so these are kind of like some model outputs from the SPM mat lab scripts um but yeah we we can definitely get and also it doesn't help that they're black and white yeah so it's a little bit like well but are we expected to I mean how can you trace so I agree they're they're not super informative there do we know at least like what they're meant to be showing like what is yeah I think like I could say the main idea if you show the diagram there if you go up to the diagram the main idea is that there's an amateur musician kind of playing some of the notes wrong and um but if you go if you go to the if you go to the screen someone's listening and they kind of know what the melody kind of should be it should just go dun dun dun dun and then D so when the musician misplays it which would be the lower right they they misplay it but the person who has the idea of what the melody should probably be like they interpret it well it's gonna it really was meant to be that it's the intended signal as opposed to the actual signal and so active inference is able to you know if you have a strong enough Matrix where the pattern is so…