Integrated Information Theory of Consciousness (IIT) Explained by Christof Koch [Interview Clip 2]

Influential consciousness researcher Christof Koch describes Integrated Information Theory (IIT). Dr. Koch is a contributor to IIT and studies consciousness at the Allen Institute for Brain Science, which he used to run.
Full interview: / ihmcurious
More clips: • Christof Koch Intervie...
For a more complete description of IIT, see the books and audiobooks below.
Video about researchers calling IIT "pseudoscience": • Consciousness Theory D...
Get Christof's books and audiobooks (I will get a small commission at no cost to you to support the channel)
- BOOK ABOUT IIT: The Feeling of Life Itself: Why Consciousness Is Widespread but Can't Be Computed amzn.to/3xiJvsd
- UPCOMING BOOK: Then I Am Myself the World: What Consciousness Is and How to Expand It (Preorder for May 7, 2024) amzn.to/3vt45FX
- EASY TO READ: Consciousness: Confessions of a Romantic Reductionist amzn.to/49oPBom
Christof's pen (MUJI pen, writes like a dream): amzn.to/4aPGvCa

[@JohnnyThomas-py3jv]

This video offers a unique perspective.

[@JediBuddhist]

Christof makes his real point 8min in. Theres a book by R D Lang called knots which gives many examples of what he is describing. There needs to be an overwhelming driving force comparing the emergent nuronal colonies. Pushing the stop feedback button and moving the system forward as a whole. Without something similar all conscious groups of neurons would be inherently mad.

[@Firestorm12345678910]

He mentions that it is not a process so how is it not a process?

[@tylermacdonald8924]

This is a huge step in the right direction

[@alexisjsosa]

Using the vast and intricate expanse of our consciousness to define itself seems fruitless. Forgive my ignorance, but has any definition or further understanding of human consciousness actually assisted with a tangible benefit - i,e. Real world applications?

[@mickmickymick6927]

Theories of Consciousness now are like the ancients discussing life-force and 5 elements. Until we have better tools which can more accurately get information on how the brain works, we're just groping in the dark.

[@iuer4643]

a few years will tell if this is useful or not, cause you can declare functions and obtain numbers, but will they be meaningful with physical sense? comes to mind the concept of enthalpy, maybe calling it "consciousness" is not quite adecuate

[@kras_mazov]

So can we build it?

[@vladyslavkorenyak872]

I don't see how it solves anything. A 1 to 1 mapping means that there is a function that produces a conscious experience for every "state" of the system. But there is no way to determine what constitutes a system. Is a metal lattice conscious? Every atom in the lattice has a causal relationship will all the other atoms in the lattice!! There is also no method to determine the particulars of this conscious experience. Could you distinguish pain from pleasure? If it doesn't answer these questions, then it is not a theory of consciousness.

[@ihmcurious]

They have ways to delineate systems, and ideas about how to derive specific phenomenal properties of experience, that we didn't have time to get into. But we may never know if they're correct. In case you want to go down a rabbit hole, Scott Aaronson made a similar critique, and Giulio Tononi wrote an interesting response: scottaaronson.blog/?p=1823

[@BladeTrain3r]

At a most basic level, I would say a system is any set of interactions with some form of entanglement relationship. If the nature of one of those interactions changes (spin changes from left to right for example), this influences the nature of the entangled interactions too. Systems can and do comprise larger systems, and as above so below. Mathematical expressions form an equation as a system, with variables, constants and operators as entities that interact to find an equilibrium. Physical objects form a system comprised of the interactions of the quantum fields comprising them. A brain is a system comprising of electrochemical neuron signals, cell activity, and resource flows. The intangibles like thought, too, are interactions in some sense - as evidenced by the fact they can drive our action. Describing such an interaction is challenging, but like many other problems we may be able to infer something by expanding on the interactions it most strongly influences. Conscious thought, as a process and core ongoing interaction in the system of a cognisant mind, can then be considered as an emergence of many smaller systems from cell signal propagation to sensory filtering to emotional processing to internal modeling, producing interactions which only obliquely exist in a physical sense, the product of it's parts. It's the "oblique" part that makes it so difficult, because it makes describing the subjective phenomena that only ephemerally arise as interactions difficult. It's hard to describe a thought quantitatively, for the same reason it's hard to describe exactly what the weightings mean in a machine learning model at any given time - billions of bespoke interactions understood only within the context of the system, almost impossible to decipher outside of that system without a full map and complete breakdown of it's components and rules.

[@tylermacdonald8924]

Well a system is itself a bit of a vague term sure, but certain collections of matter clearly have different functional relationships than atoms on a lattice structure. They may interact but it is high decoupled, and this is reflected in the degrees of freedom it has before you lose causal relationships. An emzyme in a cell on the other hand will be coupeld and reactive, and create some enduring conditions. It has a multitude of states that still allow it to return to and re-create previous states, i.e. it has much higher degrees of freedom.

[@barryrogersbarduco63]

The big question is to feel and know that you are feeling and know that you know that you are feeling and know that you know that you know that you know that you are feeling and so on to infinity.

[@StuMas]

1. For our theories to be correct, we must first accept that it is possible to attain knowledge. 2. Only that, which is discernible, is perceptible. I.e. contrasting differences helps to define the boundaries of an object, enabling it to be distinguished from its environment. We're almost blind to similarities and can only notice differences. 3. Therefore, it would follow that, anything that can be observed by a consciousness, cannot be a part of that consciousness. Because, not only would it mean they are the same thing but, the lack of a clear distinction between the observer and the observed would render the object imperceptible. I think this is the main reason why consciousness seems so fundamental yet, so elusive. Ironically, the deeper you delve into the brain in search of consciousness, the farther you'll get from finding it. Its very nature dictates it remains 'behind the scenes'. Whatever you can see through a telescope, cannot be the telescope. The same is true of consciousness. We can perceive our physical body and other's bodies but we cannot perceive perception. We cannot use light to see light. A camera cannot photograph itself directly.