Seemingly Any Computer Could Predict Elections Better Than in 2016. Here’s How Quantum Computers Might Help…
How quantum computing will be used to model elections
Excerpts and salient points ~
+ We had, in the 2016 model, for instance–regardless of what other errors might’ve come in during sampling on that–we had models that would predict a Clinton win by about nine-to-one odds. And yet, when they ran those exact same models through–which are much more difficult when you’re looking at individual states–every state has a chance to win or lose by so much. These all have to be shuffled together with individual probabilities, and we use approximations to make that easier.
+ We don’t have to do that with quantum computing. We used an [Unsupervised Deep] Learning model called a Boltzmann machine, which we probably should save for another occasion. That one predicted–using the exact same data and the same general trends–actually showed Trump more likely to win that election by about two-to-one, in some cases, but certainly not 10-to-one against.
+ The biggest thing about a quantum computer–and you were saying earlier about parallel processing. It’s not running a bunch of processes simultaneously; it’s running all of the processes simultaneously. Every single option that could be put in that, is simultaneously happening. What we do is try to remove all the ones that we can that we don’t want to see, so that we only get the actual desired one that comes through to us, when we look at it.
Content may have been edited for style and clarity.