- This event has passed.
Integrated Information Theory: Current State and Perspectives
June 24 - June 25
One of the most intriguing theories of consciousness developed during the past twenty years is the Integrated Information Theory (IIT). Its core idea is that consciousness is identical with integrated information, that is, with the amount of information a system can generate as a whole, over and above the information produced by its component parts taken independently from each other.
IIT represents an articulate, interdisciplinary attempt at characterising what consciousness is and how it works in a scientifically robust way, by making use of mathematical tools to precisely describe phenomenal consciousness both quantitatively and qualitatively, which is what makes it such a momentous theory of consciousness. However, empirically testing this theory is quite troublesome, as this would require a presently infeasible amount of computations. Furthermore, after the first phase in which the theory has gradually reached its current shape, a number of works have been published which point at a range of aspects of the theory in need of elaboration or reconsideration.
This conference is thus meant to provide a space in which some of the experts who have been contributed to IIT in various ways over the past few years will present and confront their views on the current state of IIT, on the ways it could be developed in future, and, more generally, on its prospects as a theory of consciousness. This would be beneficial for the development of IIT itself, as it would represent an important occasion to take stock and discuss the most promising directions that this important theory of consciousness may (need to) take in the future. Therefore, one of the main purposes of this conference is, crucially, to facilitate the first step towards a more mature stage of the theory, one at which, perhaps, it will become the leading contemporary theory of consciousness.
Programme (all times are in UTC+1)
Day 1 (Thursday 24th June)
14:45 – 15:00 Introduction
15:00 – 16:15 Larissa Albantakis (Wisconsin, Madison) – Integrated Information Theory – An Introduction
16:30 – 17:45 Haoying Liu (Massachusetts, Amherst) – On the Postulates of the Integrated Information Theory of Consciousness
18:00 – 19:15 Melanie Boly (Wisconsin, Madison) – Integrated Information Theory – Empirical Predictions
Day 2 (Friday 25th June)
13:30 – 14:45 Timothy Bayne (Monash) – IIT and the Unity of Consciousness
15:00 – 16:15 Adam Barrett (Sussex) – The Strength of Weak Integrated Information Theory
16:30 – 17:45 Pedro Mediano (Cambridge) – Beyond Integrated Information: Information Decomposition and its Implications for Consciousness Science
17:45 – 18:00 Concluding Remarks
If you would like to attend the conference, please register by emailing Niccolò Aimone Pisano (email@example.com) or Tom Kaspers (firstname.lastname@example.org).The conference is free and open to all.
Larissa Albantakis (Wisconsin, Madison): “Integrated Information Theory – An Introduction”
Integrated information theory (IIT), developed by Giulio Tononi and colleagues, aims to provide a theory of consciousness with explanatory, predictive, and inferential power starting from phenomenology itself. IIT contrasts with current approaches that start from the NCC with the hope of identifying generalized principles about the nature of consciousness. Instead, IIT first identifies the essential properties of every experience from consciousness itself. From these “axioms” of phenomenology, IIT then infers a set of requirements (“postulates”) for a physical system to be a substrate of consciousness. IIT predicts that the quality or content of an experience is structurally identical to the cause-effect structure of its physical substrate, and that the quantity or level of consciousness corresponds to the amount of intrinsic cause-effect power the substrate has onto itself. IIT’s theoretical formalism makes it possible to evaluate, in principle, whether a physical system complies with the IIT postulates and thus forms a physical substrate of consciousness. It, moreover, quantifies the level of consciousness of such a system, and provides a full account of its phenomenological structure in causal terms.
Haoying Liu (Massachusetts, Amherst): “On the Postulates of the Integrated Information Theory of Consciousness”
The Integrated Information Theory of Consciousness (IIT) begins with five “axioms” about the phenomenological features of consciousness. It then proceeds with five corresponding “postulates,” which are supposed to specify a system’s physical features that realize these phenomenological features. Assuming that the axioms of IIT are not false, I examine how the postulates of IIT may be derived from the axioms. Two major issues will emerge in this examination. The first concerns whether the concept of causation (in particular, cause-effect power) in the formulation of the postulates is ambiguous. The second concerns whether the postulates sufficiently capture certain robust aspects of consciousness mentioned by the axioms, such as “intrinsic perspective” and “phenomenal distinctions.” The examination will cast doubt on IIT’s potential to offer a substantive theory that sufficiently captures consciousness in terms of causation.
Melanie Boly (Wisconsin, Madison): “Integrated Information Theory – Empirical Predictions”
Timothy Bayne (Monash): “IIT and the Unity of Consciousness”
An appeal to the unity of consciousness is at the heart of the axiomatic case for IIT, but what exactly is the unity of consciousness and to what extent does it support IT? This talk considers various conceptions of the unity of consciousness, and compares and contrasts the appeal that is made to the unity of consciousness in IIT with the appeal that is made to it in other theories of consciousness.
Adam Barrett (Sussex): “The Strength of Weak Integrated Information Theory”
IIT has divided the consciousness science community. While some scientists believe it provides a feasible approach to address the `hard problem’, others dismiss it on the grounds of scarce empirical support and in-principle untestability for its primary claims. In this talk I describe how this might be ameliorated by the division of IIT into two branches, `strong’ and `weak’. Strong IIT will continue to assert that consciousness is integrated information, and search for a fundamental mathematical formula for this based on first principles. Meanwhile, weak IIT will focus on developing and testing pragmatic hypotheses on correlations between aspects of consciousness and broader measures of information dynamics, and through this strategy obtain a deeper understanding of consciousness supported by progressive empirical validations. I will go over mathematical and physical arguments for why the existing formulation of strong IIT fails, and mention some possible ways forward. I will summarise experimental evidence for weak IIT, while also discussing theoretical limitations of approaches based on one-dimensional measures of integrated information. Finally, I will argue for a multi-dimensional approach to measuring integrated information in weak IIT, and briefly touch upon the `integrated information decomposition’ taxonomy of information dynamics.
Pedro Mediano (Cambridge): “Beyond Integrated Information: Information Decomposition and its Implications for Consciousness Science”
A crucial part of the research programme in weak integrated information theory is the computational and empirical validation of practical measures of integrated information. To this end, in this talk we will first examine the behaviour of several proposed measures of integrated information, showing that they behave inconsistently even in simple linear systems. With the aim of understanding and alleviating some of the problems with these measures, we will introduce the framework of `integrated information decomposition,’ or ΦID, which allows us to decompose the `modes’ of information dynamics taking place within a system. Some of these modes will be identified with measures of integrated information, or other dynamical phenomena like causal emergence. We will then examine some of the philosophical implications of ΦID, and argue that the modes of information dynamics revealed by ΦID provide a promising avenue to explore the multi-dimensional structure of consciousness. We will conclude by outlining a few neuroscientific applications of ΦID, highlighting how a `weak’ approach to IIT can provide further insight into brain function and point towards compatibilities between IIT and other theories of consciousness.