The network brain: Seeking emergence through autonomy
February 22, 2011 Leave a comment
I’ve been coming at the problem of emergent consciousness from several angles lately. I’ve been thinking of bio-structural mimetics, perhaps. Check out these thought experiments:
The Network Brain
- A network of computers, each possessing analog inputs and outputs. Each individual computer is the analog of a neuron.
- [The inputs and outputs are the analogs of a synapse. These inputs and outputs could be biochemical, electronic, or physical but the possibility space for them should be of a similar scale as that of a synapse. By this I mean that the number of possible states should be similar to that of millions of neurochemicals interacting with hundreds of thousands of receptors in an environment of electrical charge composed of billions of charged ions. A gas chamber with special interface membranes comes to mind.]
- To begin with, the system could be set up as the virtual brain of a nematode worm (some species has a simple brain and has been extensively studied, C. elegans, I think, with 302 neurons) or a marine ragworm (has cerebral cortex).
- Using either a virtual simulation of a worm OR, excitingly, this laser control system of a live worm, the virtual brain could gain control over a living organism.
- The system would need bio-feedback. Perhaps this could be inferred from live worm observation or even directly piped into the network.
- The upshot would be a simple, proof-of-concept virtual brain. The system’s two novel components are
- a complex method of synapsing that could be extensively tuned due to its massive possibility space and
- a method of interacting with a biological system. The biological system could at first be a read-only depiction of the brain-network’s states, but could be enhanced to provide sensory feedback, whereby the virtual brain would ‘possess’ the biological host. This host’s interaction with biological organisms could form the first instance of neosentient empathy.
- This program lays a foundation for modeling systems of distributed computation using networks of binary computers. Each computer in the network should be able to provide low-latency processing of inputs from the physical interface with other members of the network and return outputs, enabling deep mutual entrainment, propagation of synapse waves, and other characteristics of neuronal systems.
- Wolfram’s automata (particularly those of type 3 or 4 complexity) exhibit amazing complexity from simple rules. They could form a more manageable way of generating emergent complexity (presumably, a phenomenon related to consciousness) than the network system described above
- again individual CPUS (or, possibly, separate programs on a single machine) would interact. Wolfram’s automata may offer the needed complexity and expansive possibility space of the physical interface above without the overhead.
The simplicity of consciouness
- As in the article described here, there is mounting evidence that the cortex developed relatively early on in animals- and it is relatively simple. This indicates that we might be able to develop whatever sentience is entailed by the cortex in a simple system as described above.
Other thought experiments to follow.