Questa pagina è stata tradotta automaticamente dall'inglese. Visualizza l'originale in inglese.

Parr, Pezzulo, Friston 2022 Textbook Cohort 6, Chapter 3, part 1

Textbook Group meeting for Parr, Pezzulo, Friston 2022 .

Mar 4, 2024

▶ Watch on YouTube ↗

Session details

Date: Mar 4, 2024

Series: Parr, Pezzulo, Friston 2022 Textbook Cohort 6, Chapter 3, part 1

Transcript

AI-generated transcript excerpt

The full transcript is available on GitHub. This excerpt is generated by automated speech recognition and may contain errors.

all right welcome back it's March 4th the only day that's a command March 4th so as we head into chapter 3 does anyone want to begin with any any thoughts or questions or remarks about three or where is it in the context of the book what do they remember from it or what's a quote that was interesting or just a point to jump in from I personally found it really interesting that the marov blanket is just an intermediary it's just a barrier between two things the external and the internal State I really it reminded me of my physics classes talking about a ball rolling down the hill in the sense that you could take it from the perspective of the ball and the universe of moving around it or the the universe is still and the ball is moving down the hill that freedom for that kind of reference surprised me yeah interesting stable Hill moving ball and stable Ball moving Hill that that is that it Zach absolutely yeah yeah that a kind of similar topic came up yesterday in the um discussion with llan Kent and Anna Unica on mental gravity on like kind of there's a reference frame where the Earth like spaceship earth is the egocentric navigation that's like our visual reality like objects coming closer and then there's kind of a third person Sun centrism there's other reference points that can be chosen but that's definitely a very so how would you connect that to markup blanket or to AC or how does the Markov blanket allow us to to do it this way wait I'm sorry perhaps I don't understand are you saying they posed that question or are you asking that question oh just asking like what what made you what made you think about this ball moving down the hill from the markof blanket oh because of the K Divergence it's not exactly perfect but it seems as if you uh can move X the actual like external world is going to stay still but the Q is going to be the thing that's changing however it seems to where your choice as the ball as the thing moving down the hill is kind of arbitrary if you assume that the K Divergence is not going to change if you change up the order it was just a thought that I had that found to be interesting the kind of relative relationship between the two perhaps there's not much to it well well I'm I'm seeing maybe the lrange points as being somewhat porary what are those points so when you have an orbit um when you have two bodies in orbits there are certain points where there's an equilibrium between the gravity Fields between those two bodies so the lrange point is actually a place of stasis um and in a way because it's an absence of both those bodies and the in the influence of both those bodies it's actually a separate virtual body in and of itself so or a virtual Point excuse me that that's very interesting lock Li in in this kind of mental gravity um discussion which maybe people can look at for more detail it's like he did discuss the center of gravity those topics came up with the basian mechanics and Dalton otiva devel like you have a distribution and then you have a center of Mass center of gravity that's like the central tendency of a distribution so basing mechanics is kind of something like a gravitational interpretation of the basian updating process like big distribution heavy prior Light low attention uh sensory input not going to move the prior that much and then the um lrange point is kind of like the the stationarity if you're there like a non-equilibrium stationary Point like uh the person in the frog and the uh jumping out of the hand example a textbook they could reach like a long-term stationerity where they were just estimating something that where they kind of they've reached like a steady point with what their environment was providing so it's kind of like one Gravity from the environment one Gravity from the um agent with their prior and then also that the um p is stable like the true distribution of uh temperature stable Q is changing beliefs are changing or beliefs are stable true distributions…