Archive
Weekly QuEST Discussion Topics, 21 May
QuEST 21 May 2021
We will start this week by beginning a series of short talks from the QuEST group that is writing articles on the QuEST Theory of Consciousness.
The QuEST Theory of Consciousness (S3Q, https://arxiv.org/ftp/arxiv/papers/2103/2103.12638.pdf) suggests all aspects of the conscious representation are defined by relationships and interactions, i.e., the situatedness of the representation. We will present a proposed hierarchical system of brain rhythms to ground S3Q in modern neuroscience (https://www.sciencedirect.com/science/article/pii/S0896627313009045#fig1). We will then discuss how this framework can be used to investigate the relationships composing conscious representations in humans, nonhuman animals, and machines (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2605405/).
If time permits then want to continue to talk this week about the Free Energy Principle and related topics like our Representational Correlates of Consciousness and insights from computational correlates of consciousness to the RCC idea. In addition to continuing the discuss the Wiese / Friston paper:
The neural correlates of consciousness under the free energy principle: From computational correlates to computational explanation
Wanja Wiese (wawiese@uni-mainz.de), Karl J. Friston (k.friston@ucl.ac.uk<mailto:k.friston@ucl.ac.uk>)
* How can the free energy principle contribute to research on neural correlates of consciousness, and to the scientific study of consciousness more generally? Under the free energy principle, neural correlates should be defined in terms of neural dynamics, not neural states, and should be complemented by the research on computational correlates ** recall QuEST RCC – representational correlates of consciousness focus of consciousness – defined in terms of probabilities encoded by neural states.
* We argue that these restrictions brighten the prospects of a computational explanation of consciousness, by addressing two central problems.
* The first is to account for consciousness in the absence of sensory stimulation and behaviour.
* The second is to allow for the possibility of systems that implement computations associated with consciousness, without being conscious, which requires differentiating between computational systems that merely simulate conscious beings and computational systems that are conscious in and of themselves.
We will add
Learning to Be Conscious
Axel Cleeremans,1
112 Trends in Cognitive Sciences, February 2020, Vol. 24, No. 2 https://doi.org/10.1016/j.tics.2019.11.011
(c) 2019 Elsevier Ltd. All rights reserved
* Consciousness remains a formidable challenge. Different theories of consciousness have proposed vastly different mechanisms to account for phenomenal experience.
* Here, appealing to aspects of global workspace theory, higher-order theories, social theories, and predictive processing, we introduce a novel framework: the self-organizing metarerpresentational account (SOMA), in which consciousness is viewed as something that the brain learns to do. By this account, the brain continuously and unconsciously learns to redescribe its own activity to itself, so developing systems of metarepresentations that qualify target first-order representations. Thus, experiences only occur in experiencers that have learned to know they possess certain first-order states and that have learned to care more about certain states than about others. In this sense, consciousness is the brain’s (unconscious, embodied, enactive, nonconceptual) theory about itself.
Weekly QuEST Discussion Topics, 14 May
QuEST 14 May 2021
We want to talk this week about the Free Energy Principle:
The neural correlates of consciousness under the free energy principle: From computational correlates to computational explanation
Wanja Wiese (wawiese@uni-mainz.de), Karl J. Friston (k.friston@ucl.ac.uk<mailto:k.friston@ucl.ac.uk>)
* How can the free energy principle contribute to research on neural correlates of consciousness, and to the scientific study of consciousness more generally? Under the free energy principle, neural correlates should be defined in terms of neural dynamics, not neural states, and should be complemented by the research on computational correlates ** recall QuEST RCC – representational correlates of consciousness focus of consciousness – defined in terms of probabilities encoded by neural states.
* We argue that these restrictions brighten the prospects of a computational explanation of consciousness, by addressing two central problems.
* The first is to account for consciousness in the absence of sensory stimulation and behaviour.
* The second is to allow for the possibility of systems that implement computations associated with consciousness, without being conscious, which requires differentiating between computational systems that merely simulate conscious beings and computational systems that are conscious in and of themselves.
Weekly QuEST Discussion Topics, 7 May
QuEST 7 May 2021
We have a couple of topics trending this week. Feel free to chime in with your own.
Narratives – possibly discuss our internal work attempting to map the QuEST definitions / tenets onto a policy gradient reinforcement learning solution.
AI and education for the inner city. Several of us have been having discussion with people interested in the use of AI to help student regain what they lost in the Covid year out of the classroom. To that end we’ve been looking at different existing AI for education commercial solutions.
In a recent podcast from the Machine Learning street talk, #52, with Hadi Salmon (PhD student at MIT) on unadversarial examples. We’ve discussed adversarial examples on several occasions but recent work that attempt to use the insights from adversarial examples to parse ‘robust’ versus ‘nonrobust’ features and then measure the impact on transfer learning.
Melanie Mitchell recently wrote an article listing 4 fallacies that make AI harder than we think:
Quest Group's Blog 