Communications of the ACM

Unite neuroscience, supercomputing, and nanotechnology to discover, demonstrate, and deliver the brain's core algorithms.

Anonymous

Conggratulations first. Then disappointment (as with Watson originally designed to do translations (speech to speech) and now operating a call centre.) the reason being that the biophysical structure, mechanism and function of the brain is just an imprint of the processes going on in the brain, a very little part of which is available to the conscious mind, hence may be righfuly called as cognition. Doing away with von Neuman's machine is a good step, getting rid of Greek grammar, philosophy and formal logic should also follow. Be reminded that information is always recorded on a surface, and the surface is formed by opposing forces (energy in wave forms) all synchronised, obeying the principles of symmetry. But as with the cycles of water, not everything is pushed and pulled by the local factors here on the earth only. Keep up the good work and engage some out of the box thinking linguists as his paradigm may come useful.
F. Kovacs

Anonymous

I thought this article was important in that it emphasized, in a way, the sheer size and magnitude of the effort to simulate a human brain's cognitive structure. The writing was very succinct and formal at times, but appropriate considering the scientific material. I just want to say thanks for featuring this article, it was fascinating. Hopefully it is not too optimistic to imagine a supercomputer simulation of the human brain, as stated at the conclusion, within the next decade! Very optimistic, indeed. Let's just hope the "business machines" envisioned as the upshot of this work, won't simply function as another way to eliminate humans from the workplace. Kudos to ACM for the selection of the subject matter, and kudos to the authors of the article.

Anonymous

Excellent work, I'm very excited to see what this project can bring about.

Thank you for sharing all these insights and best of luck in unlocking the mysteries of the mammalian mind.

D Molnar

Anonymous

Consciousness is a direct result of building a brain for predicting what other actors will do, what they are interested in, what are their capabilities and tendencies. To succeed as a tribe member we each have to understand how to capitalize on the tribe's power. That means simulating power broker reactions(starting with our parents and siblings). The decision simulator can then be used on our self. It reports how we believe we will react. Which is why our accuracy is so bad as the simulator has to compete with more primal drivers who are more connected with reactions than the simulator/reporter/analyzer loop which is looking for a high value result which may be hidden in the details.

Anonymous

I really do not understand what this project will bring beyond the capabilities of current AI approaches. Authors say that current AI developed unique approaches for unique problems, that is right because this is what human do, they develop unique approaches because they need unique tasks to be performed. Generalizing AI approaches does not mean creating new discipline of science. However, i hope to see the outcome of this project

dan tso

The "essentially digital electrical signal of the spike" is an all too common characterization of neural signaling down an axon. It is more appropriate to instead say that the analog neural signal in the amplitude (voltage/current) domain has been traded for an analog neural signal in the time/frequency domain for the purposes of accurate transmission down leaky, lossy axons.