Thoughtworks Ignite On Brains And Computing
2 likes655 views
1) The brain is a complex network of neurons and synapses that functions in a massively parallel and distributed manner. 2) Individual neurons perform simple computations by transmitting electrical signals via ion concentration gradients, while synapses vary in efficacy to perform "weighting" between neurons. 3) Neuromorphic engineering aims to understand and emulate the brain's principles in hardware to further research and potentially develop useful applications.
Thoughtworks Ignite On Brains And Computing
- 1. On Brains and Computing Anthony Hsiao SapnaSolutions
- 2. About Me
- 3. „ “ The Brain – My second most favourite organ! (Woody Allen, Director)
- 4. Context Neuroscience is the scientific study of the nervous system Informatics is the science of information, the practice of information processing, and the engineering of information systems Neuroiformatics Neuromorphic Engineering
- 5. The Brain in a Nutshell Complex network of Neurons Synapses Distributed Brain Sensors CNS Stomach
- 6. Neurons – Computation Concentration gradient of Ions K+, Na+,Ca2+ Action Potential Electrical Performs 'Sum'
- 7. Synapses - Memory? 'Gap' between two interconnecting Neurons Transmits AP Varying Efficacy Action Potential Chemical Performs 'Weighting'
- 8. The Amazing Brain Simple repeated pattern: Concentration difference of Ions in a liquid Many Neurons > Many other Neurons Giving rise to: Emotions Thought Sensation Realities
- 9. Quick Comparison: Brain and Computers Brain Computer Distributed, Centralised, still Architecture massively parallel, mostly serial redundant Operating Freq. 50-100Hz > Ghz Power Consump. ~20W 150-300W 10e11 Neurons, Complexity 500M 10e15 Synapses Processing PS3 Cell – 10k ~ 100M MIPS Power MIPS Feature Size 4-100um ~45nm Memory 100TB a lot as well ...
- 10. Neuromorphic Engineering Understanding the principles by which brains function Emulate those in hardware, for further reasearch, and ultimately, to find useful applications Most Neuromorphic hardware is Analog!
- 11. Neural Circuits
- 12. Science today?
- 13. Who is making decisions? Decisions 'made' in the brain before you know it Subjects asked to press a button Scanner revealed response (Haynes 2008) M a x P la n c k In s titu te fo r H u m a n C o g n itiv e a n d B r a in S c ie n c e s in L e ip z ig , C h a r ité U n iv e r s ity H o s p ita l a n d th e B e r n s te in C e n te r fo r C o m p u ta tio n a l N e u r o s c ie n c e in B e r lin
- 14. Cool Stuff
- 15. More cool stuff
- 18. Can we build intelligence? When will computational power of machines match that of a human? When will computational power of machines match that of humanity?
- 19. Can we build intelligence? When will computational power of machines match that of a human? > 2029 (Kurzweil, passing the turing test) = 1 HB (Human Brain) When will computational power of machines match that of humanity? > 2050 (Kurzweil, Kelly) = 6bn HB
- 20. What is free will? If our experience of reality is generated by our brain, given the research I showed you, how do we explain that our impression of choice seems to stem from an underlying mechanistic structure?
- 21. Can machines feel? Conversely, if despite this mechanistic nature of our brain, we are indeed emotional, then can't sufficiently machines and robots be emotional too?
- 22. Final thought Who can think of a similar man-made structure that has some resemblance to the brain?