Sciopartum
A working theory of consciousness based on research in computational neuroscience and machine learning
…
Consciousness research has reached an inflection point, transitioning from philosophy to the hard sciences. Though self-introspection will always remain an important part of consciousness research, we believe the best theory of consciousness should be guided primarily by modern neuroscience and statistics to tackle the hardest problems in philosophy of mind head on: The Hard Problem, The Binding Problem and The Boundary Problem.
“Sciopartum”, also referred to as “Sciops Theory” represents a working draft of a theory of consciousness that we feel best matches the scientific research conducted by neuroscience and machine learning labs at this moment in time. As such, the ideas expressed here are subject to change as new experimental evidence comes in, or as new statistical analysis are conducted.
Though a detailed read of the chapters is necessary to grasp the entire theory, a few key concepts are worth mentioning here to give a high level overview of the theory as it stands.
Thus far, we see no scientific evidence for many of the common assumptions historically made by the philosophy world regarding consciousness. For example we see no evidence for a binary or harsh cut off between computational systems that are conscious and those that are not. Instead, the statistics point to the idea that our conscious experience is bound by the convergence of information in the brain (or any computational system) and is limited by the relative statistical uncertainties of each piece of information. Thus early evidence seems to suggest that simple bayesian probabilities shape the boundary of consciousness, not some sort of secret ingredient. Simply put: we experience the information that the highly integrated areas of our brain actually have access to when making key decisions, analyzing our environment, and taking actions. Further research via neural decoding should shed more light on this issue.
Similarly, we also see no evidence of a secret ingredient needed to bridge The Hard Problem, as our early analysis of the mathematical structures of neural networks seem to give rise to logical structures which could plausibly one for one correlate with key phenomenal experiences such as visual space, color, sound, touch and even joy/pain. Though further research needs to be conducted to see if these mathematical structures can fully explain our qualia, we see no reason as of yet to assume this is implausible, or that some missing ingredient is needed to fully bridge the gap.
By further pursuing scientific research down several promising avenues (for example: connecting neural decoding tasks with conscious experiences, deepening statistical models of information in the brain, conducting sensory experiments, etc) we feel there is a strong likelihood that in the future we will be able to definitively answer existential questions such as…
- Is AI conscious? If so, what is their experience?
- How do we create new conscious experiences or manipulate our experiences?
- What is the mathematical structure for vision? Sound? Joy?
- How do we understand personal identity? And what happens after we die?
— — —