/sci/ - Science & Math

http://news.mit.edu/2016/method-image-black-holes-0606
Last paragraph.
>The authors of this paper use a highly advanced approach to learn prior knowledge,” he continues. “The application of this prior-model approach to event-horizon images is not trivial. The authors took on major effort and risk. They mathematically merge into a single optimization formulation a very different, complex sensing process and a learning-based image-prior model.

The "very different" complex sensing model was devised in the 50's
https://en.wikipedia.org/wiki/Very-long-baseline_interferometry
https://en.wikipedia.org/wiki/Closure_phase
https://en.wikipedia.org/wiki/Roger_Clifton_Jennison

The only new thing they seem to have done is feeding their simulations of black holes into their AI in order to train it to do a "better" reconstruction.
The problem is that is the most dubious part of the whole fiasco.
The construction of a better prior relies on assuming current physics is correct and that the simulations are accurate/representative.

>Assume you know what you should see
>See what you assumed you should see

The ABSOLUTE state of science.