This Thursday at 6:30pm, join us on Zoom for our first student talk this year with our friends at the Theoretical Physics Student Association! You can find the link for it in the weekly email!

Michael Mitchell talks about his summer project concerning image processing for certain experimental situations which must contend with a low amount of signal present. Such as electron microscopy, where high energy beams would result in specimen damage. Low signal incurs Poisson noise, less useful data, and difficulty focusing and tuning the experimental setup.

Using a Hybrid Machine Learning algorithm architecture to enhance this data in real-time aids the experimentalist, all while mitigating the introduction of non-physical artefacts from the training set in a mathematically demonstrable way.