Mining algorithm based on neural networks

[nickmiller24]

One advantage of GPUs and (probably) FPGAs is they can be re-purposed to do things unrelated to the cryptocurrency space. GPUs can also be used by gamers and anyone who needs a lot of computing power. And FPGAs can be reprogrammed to do almost any specialized computing task.

ASIC machines could also theoretically be repurposed to perform another task besides mining as long as the other task used the same algorithm they were originally designed for. For example bitcoin ASICs could potentially be repurposed to perform other tasks that require the SHA256 algorithm or similar mathematics. There is not a huge amount of people who want to perform SHA256 really fast, thus there is not a huge secondary market for bitcoin ASICs. This could cause waste in the future because if there is no secondary market and mining becomes unprofitable for any reason, these ASICs turn into paperweights.

To avoid paperweights being created coins should be using a cryptographic hash algorithm that has some kind of secondary demand. One way to do this is by using a neural net as a cryptographic hash algorithm instead of something like SHA256. The ASICs created to mine this neural net algorithm could then be repurposed to do tasks related to artificial intelligence.

This way miners could sell their specialized hardware to AI people and vice versa.

There has already been some preliminary research into using neural networks as hash algorithms, see the links below.

https://arxiv.org/ftp/arxiv/papers/0707/0707.4032.pdf

https://pdfs.semanticscholar.org/d6d0/caf56398db9679690506af5eff387e743220.pdf

https://en.wikipedia.org/wiki/Neural_cryptography

[Max Likelihood]

I am a statistician, and neural nets are in essence nonlinear regression models that predict some target variable Y based on a complex function of several input variables, say a vector X. So they are usually used to identify systematic relationships between some target and many possible predictors.

These papers describe a creative but non-standard use that I'm not sure has much use in POW mining. In mining, Y is a randomly generated target number that has to be hashed with some function, with solutions reflecting randomly generated numbers hashed through the same function then compared to see if they are < Y. There is nothing particularly special about using neural nets for the hashing, because one is just trying still by luck to obtain a solution smaller than the target. There are no systematic regression relationships among variables to be modeled with the nnet. If there were systematic predictors of the target it would actually be much easier to hit it.

In proof of useful work, real models with real data might be substituted, but it is a different issue than using nnets to construct a hash function which seems computationally wasteful.