4 ms·

The article seems a bit abstract. I think the question that comes to mind is whether it's possible to take GPU-styles architecture and give it 100x more power

The article seems a bit abstract. I think the question that comes to mind is whether it's possible to take GPU-styles architecture and give it 100x more power or more by replacing bits with approximate voltage levels that are "fuzzy" but statistical guarantees to their performance, along with gates that allow the values to be filtered back to zeros and ones. The thing is the "neural architectures" seem to have been more or less failure through requiring the user to accept one particular neural net structure while the standard GPU seems to have succeeded through being generic-enough to use for a variety of tasks. So some sort of analogy-GPU should also have a similar generic model - but of course it seems likely that the creator of this stuff will want to impose their special model. Edit: More or less like the D-wave "quantum computer" except not milking quantum hype and being reconciled to being understood as "massively analogue"

krastanov 8y ago
The "fuzzy but statistical guarantees" is basically "error correction". Which is (handwavily) the difference between digital and analog computing. What you are describing is the engineer's definition of a digital computer. Admittedly, there might be interesting performance gains if we use less stringent statistical guarantees... which happens to be what is happening each time we make the transistors smaller and more susceptible to noise. C.f. the paper that introduced the distinction between analog and "statistically guaranteed" digital in the case of classical computers (before it people were arguing that you can not build a scallable classical digital computer because of noise): by von Neumann http://www.sns.ias.edu/pitp2/2012files/Probabilistic_Logics.pdf http://www.sns.ias.edu/pitp2/2012files/Probabilistic_Logics.... The paper that did the same for quantum computers 50 years later: by Shor http://www-math.mit.edu/~shor/papers/good-codes.pdf http://www-math.mit.edu/~shor/papers/good-codes.pdf P.S. FYI DWave is not a scallable quantum computer. They have another "quantumy" word for what they do, so that they can keep the hype without angering people that are trying to build an actual quantum computer.
- kangnkodos 8y ago
  Yes. Concentrating on error correction reveals the difference between digital and analog computing. Digital computers have error correction. Analog computers don't. Some small number of problems might be solved efficiently using analog computers, but they will never take over the role of general purpose computers because of error correction issues.

krastanov 8y ago

The "fuzzy but statistical guarantees" is basically "error correction". Which is (handwavily) the difference between digital and analog computing. What you are describing is the engineer's definition of a digital computer. Admittedly, there might be interesting performance gains if we use less stringent statistical guarantees... which happens to be what is happening each time we make the transistors smaller and more susceptible to noise. C.f. the paper that introduced the distinction between analog and "statistically guaranteed" digital in the case of classical computers (before it people were arguing that you can not build a scallable classical digital computer because of noise): by von Neumann http://www.sns.ias.edu/pitp2/2012files/Probabilistic_Logics.pdf http://www.sns.ias.edu/pitp2/2012files/Probabilistic_Logics.... The paper that did the same for quantum computers 50 years later: by Shor http://www-math.mit.edu/~shor/papers/good-codes.pdf http://www-math.mit.edu/~shor/papers/good-codes.pdf P.S. FYI DWave is not a scallable quantum computer. They have another "quantumy" word for what they do, so that they can keep the hype without angering people that are trying to build an actual quantum computer.

kangnkodos 8y ago
Yes. Concentrating on error correction reveals the difference between digital and analog computing. Digital computers have error correction. Analog computers don't. Some small number of problems might be solved efficiently using analog computers, but they will never take over the role of general purpose computers because of error correction issues.