Whenever I see “machine-learning” in the context of scientific development I just mentally replace it with “magic”.
Employing machine learning, at the data acquisition stage, is a really bad idea. It can easily introduce biases...
In that hypothetical setup, we end up storing every raw datapoint. If we think of these as parameters, then PCA is a non-parametric model; although it's maximally inefficient in terms of information/compression.
This is all fine. However, the problem is suggesting that this is a somehow reasonable/tenable solution to solving the noise issues...
Maybe the noise can be removed by other methods, maybe the noise doesn’t matter, but magic doesn’t solve the noise issues.
The noise definitely matters. When you are observing with a telescope you are almost by definition working at the limit of the signal to noise ratio of the instrument, otherwise you would be asked to use a less powerful instrument (since the observing time on these larger instruments is so precious).
I worked on a project analyzing data from an X-ray telescope back in my research days. We spent 3 full years building an incredibly detailed model of the instrument noise, just so we could subtract it out and have confidence in our remaining signal. That was longer than the time we spent doing the actual science on the corrected data.
> "Analyses that typically take weeks to months to complete, that require the input of experts and that are computationally demanding, can be done by neural nets within a fraction of a second, in a fully automated way and, in principle, on a cell phone’s computer chip,” says postdoctoral fellow Laurence Perreault Levasseur, a co-author of a study published today in Nature."
> "The team at the Kavli Institute for Particle Astrophysics and Cosmology (KIPAC), a joint institute of SLAC and Stanford, used neural networks to analyze images of strong gravitational lensing, where the image of a faraway galaxy is multiplied and distorted into rings and arcs by the gravity of a massive object, such as a galaxy cluster, that’s closer to us. The distortions provide important clues about how mass is distributed in space and how that distribution changes over time – properties linked to invisible dark matter that makes up 85 percent of all matter in the universe and to dark energy that’s accelerating the expansion of the universe."
> "Until now this type of analysis has been a tedious process that involves comparing actual images of lenses with a large number of computer simulations of mathematical lensing models. This can take weeks to months for a single lens."
> "But with the neural networks, the researchers were able to do the same analysis in a few seconds, which they demonstrated using real images from NASA’s Hubble Space Telescope and simulated ones."
[1] https://www.symmetrymagazine.org/article/neural-networks-mee...
Just like parametric statistics, then?