Double Slit Experiment and Bayes (2019) | Hacker News Reader

Double Slit Experiment and Bayes (2019)

yulingyao.com

53 pointsbookofjoe8 comments

That is an extremely interesting blogpost! Thank you for sharing! It's been a problem I have been struggeling with lately, as to how qm seems to have negative probabilities in some contexts. I will study this blogpost deeply.

It does seem very technical for a hn audience.

Somebody could make a little bit more verbose version of this with a few pictures.

> qm seems to have negative probabilities in some contexts

This is actually how QM is described to people who study probability theory: as the extension of probability theory to include negative probabilities.

But that does not make a whole lot of sense though. What does it mean to have negative probability? Can you sample from negative probability?

Also on a more fundamental level. If you define probability coming from the idea of Dutch books (which is how I think about probabilities), then it is really easy to show that there should be no such thing as a negative probability.

I like the description in the article a lot better than the idea of 'magical' negative probabilities.

Negative probabilities interestingly are showing up in various other realms. There seems to be converging interest in the topic from a number of areas.

For example, in relating Bayesian inference to algorithmic complexity-based inference: https://arxiv.org/abs/1401.7116

I've also read some interpretations of negative probabilities in terms of erroneous values. For example, if you're sampling observations, and a machine error causes an anomalous value to appear in the data. This I don't know I completely intuit but there are interpretations of negative probabilities outside of QM.

My sense is that they will eventually be accepted akin to imaginary numbers or something but that's just a hunch and I could be wrong.

Interesting. My reaction was the opposite: I find it a very muddled and unclear blogpost, and frankly I am not sure there is anything insightful or useful in it. I’d say, yes, Answer 1 is it, you are looking at two different settings, each of which is reasonably well understood by now. What has Bayes got to do with it?

I agree, the blogpost is not written clearly. But there is something fundementally odd about how wavefunctions cause seemingly negative probabilities. And understanding that might lead to improvements in how we deal with probabilities and statistics.

Contrary to what people might think from their schoolbooks, a lot of statistics is built on a swampy foundation.

Constructive and destructive interference are generally well-understood and common phenomena even in the macro world. Sound waves, ocean waves, string instruments. With QM, the probability can never be negative since it is calculated as the amplitude squared.

Somewhat similarly, with ocean waves, we are not concerned that troughs are somehow creating negative energy when energy of an ocean wave = amplitude squared.

Reply on news.ycombinator.com