From generative models to generative passages: A computational approach to (neuro)phenomenology

Abstract

This paper presents a version of neurophenomenology based on generative modelling techniques developed in computational neuroscience and biology. We call this approach computational phenomenology because it applies methods originally developed in computational modelling to phenomenology. The first section presents a brief review of the project to naturalize phenomenology. The second section presents and evaluates philosophical objections to that project, and situates our project with respect to these projects. The third section reviews the generative modelling framework. The following section presents our new approach to neurophenomenology based on generative modelling. We then discuss how this application of generative modelling differs from previous attempts to use it to explain consciousness. In summary, generative modelling allows us to construct a computational model of the inferential or interpretive process that best explain this or that kind of lived experience.

What Might Interoceptive Inference Reveal about Consciousness?

“…In addition to this, it is theorized that consciousness is the felt affect that results from explicitly evaluating the expected free energy under different actions, as opposed to automatic or reflexive behaviour [115]. There is also an attempt by Ramstead et al [116] to apply generative modelling to understand phenomenology on its own terms, arguing that raw sensory experience can be likened to the observations of an active inference agent, and that the coherent lived experience is then the most likely posterior belief or the best explanation for those raw experiences. Many of these approaches are probably consistent or at least overlap with predictive instantiations of HOT and GNWS theories for consciousness in the brain, for they also emphasize hierarchically structured predictions of the consequences ofand the accuracy of -own beliefs.…”

What might interoceptive inference reveal about consciousness?

“…In addition to this, it is theorized that consciousness is the felt affect that results from explicitly evaluating the expected free energy under different actions, as opposed to automatic or reflexive behaviour [113]. There is also an attempt by Ramstead et al [114] to apply generative modelling to understand phenomenology on its own terms, arguing that raw sensory experience can be likened to the observations of an active inference agent, and that the coherent lived experience is then the most likely posterior belief or the best explanation for those raw experiences. Many of these approaches are probably consistent or at least overlap with predictive instantiations of HOT and GNWS theories for consciousness in the brain, for they also emphasize hierarchically structured predictions of the consequences ofand the accuracy of -own beliefs.…”

Conscious Self-Evidencing

“…In phenomenological and physiological traditions, there is debate about how to frontload phenomenology into the explanatory project and what it would mean for there to be explanatorily substantial isomorphism between physiology and phenomenology (for discussion, drawing lines back to Kurt Koffka and Edwin Boring, see Feest 2021). An ongoing project here is to more closely marry the mechanistic and computational debates with the phenomenological debates, and thereby lay the ground for scientific projects that forge the explanatorily link for measurable properties of phenomenology (Ramstead, Hesp et al 2021); a concrete example here could be the mentioned active inference account on Troxler fading and binocular rivalry (Parr, Corcoran et al 2019).…”