Substrate-Independent        Minds     Randal A. Koene     Carboncopies.org                        1
DYSTOPIA           2
SELECTION            3
EMPATHY          4
CHALLENGES             5
AUGMENTED            6
BEING        7
SUBSTRATE-INDEPENDENTMINDS                 WHOLE                  BRAIN              EMULATION8
SYSTEM IDENTIFICATION
ROADMAPREQUIREMENTS               11
SCOPERESOLUTION             12
EMULATIONPLATFORM    80 million ATP per action            potential            Brain: 20-40W (20-44% of            body)  ...
STRUCTURALCONNECTOME             14
15
16
FUNCTIONALCHARACTERIZATION                   17
MOLECULARTICKER TAPE      Kording (Northwestern), Boyden (MIT), Church (Harvard), Koene   18
FUNCTIONALIZEDNANOPARTICLES
WIRELESSMICROPROBES              Gomez-Martinez et al. (2009)                                       20
21
TEAMNETWORK          2um   1um                      22
Carboncopies.org
BBODY           BRAIN
Thank You    carboncopies.orgrandal.a.koene@carboncopies.org                                  25
Dr. Randal Koene: Substrate-Independent Minds (Australian Singularity Summit 2012)
Upcoming SlideShare
Loading in …5
×

Dr. Randal Koene: Substrate-Independent Minds (Australian Singularity Summit 2012)

1,202 views

Published on

The first keynote by Dr. Randal Koene at the 2012 Australian Singularity Summit. Dr. Koene presents the concept of substrate-independent minds (SIM) and gives an outline of a feasible approach to SIM by means of whole brain emulation.

Published in: Technology
0 Comments
2 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
1,202
On SlideShare
0
From Embeds
0
Number of Embeds
296
Actions
Shares
0
Downloads
1
Comments
0
Likes
2
Embeds 0
No embeds

No notes for slide
  • General concept of demux tree Specific implementation via nanowires, Rudolpho Llinas NYU School of Med, 500nm polymer or platinum wires
  • General concept of demux tree Specific implementation via nanowires, Rudolpho Llinas NYU School of Med, 500nm polymer or platinum wires
  • You choose a modeling resolution At it and below, functional characterization is key; the components must be simple enough to capture all their relevant behavior Above it, structural characterization is key; the interaction of components and their emergent product; connectome Function-structure in neuronal networks, and hardware/software – both are actually very similar in any other complex system, including computers You can use concurrent functional characterization and low-res structural information (e.g. finding all neurons) to infer functional connectivity – but there is risk of missing latent functions You can reduce the signal dimensions – considering only spatial architectural information (morphology) instead of spatial and temporal information (responses) by combining high-res structural information with inference of “functional type” - but even with a complete library of types and their paramter-related distributions of behaviors, there there is risk that mapping may not be one-to-one and that spatial data errors (e.g. misreadings of component sizes) may be cumulative and may be hard to correct without the addition of locally gathered functional information It is also better in general to rebuild a complex system bit by bit, while carrying out validation that it still works
  • Red blood cell diameter 8um. Existing biopassive & bioactive coatings for neural implants. Passes through every capillary to every neuron in the brain. Chip in cell has been done Artificial red blood cell has been done
  • Teams: Hub cell (8um), sensor, stimulation, herding, chains, tag delivery, morphology recording agents (1-2um)
  • Dr. Randal Koene: Substrate-Independent Minds (Australian Singularity Summit 2012)

    1. 1. Substrate-Independent Minds Randal A. Koene Carboncopies.org 1
    2. 2. DYSTOPIA 2
    3. 3. SELECTION 3
    4. 4. EMPATHY 4
    5. 5. CHALLENGES 5
    6. 6. AUGMENTED 6
    7. 7. BEING 7
    8. 8. SUBSTRATE-INDEPENDENTMINDS WHOLE BRAIN EMULATION8
    9. 9. SYSTEM IDENTIFICATION
    10. 10. ROADMAPREQUIREMENTS 11
    11. 11. SCOPERESOLUTION 12
    12. 12. EMULATIONPLATFORM 80 million ATP per action potential Brain: 20-40W (20-44% of body) Average action potential takes 1.5ms About 4nW per event Supports 7 billion concurrent events Average rate 4-7Hz... 470 billion events/s 13
    13. 13. STRUCTURALCONNECTOME 14
    14. 14. 15
    15. 15. 16
    16. 16. FUNCTIONALCHARACTERIZATION 17
    17. 17. MOLECULARTICKER TAPE Kording (Northwestern), Boyden (MIT), Church (Harvard), Koene 18
    18. 18. FUNCTIONALIZEDNANOPARTICLES
    19. 19. WIRELESSMICROPROBES Gomez-Martinez et al. (2009) 20
    20. 20. 21
    21. 21. TEAMNETWORK 2um 1um 22
    22. 22. Carboncopies.org
    23. 23. BBODY BRAIN
    24. 24. Thank You carboncopies.orgrandal.a.koene@carboncopies.org 25

    ×