If you are attending the Optica Biophotonics conference in San Diego, be sure to check out Magdalena's talk on this work!

DeepPD obtains better de-aberrated images without requiring optical correction, and more completely estimates wavefronts by accounting for nonlinearities in our deformable mirror and other assumptions not captured in our original analytic approach.

In our new preprint, we combine a deep learning framework with our phase-diverse image acquisition called "DeepPD".

We demonstrated that by acquiring few extra images with added optical aberrations ("phase diversities") wavefront sensing can be achieved using an analytic optimization scheme.

Last year we published a simple experimental method for image-based wavefront sensing in fluorescence microscopy using phase diversity.

opg.optica.org/optica/fullt...

Too bad it will involve actual legwork to distinguish against the broader uses of these terms. My research project is literally called "Phase Diversity" - it originated in the 1980's in Astronomy - I study physics, not DEI. Wait until they find out how much cis/trans gets used in Chemistry!

This is utterly insane. I literally study a method called "Phase Diversity" that originates in Astronomy and deals with physics - not DEI. I am lucky that I will not be affected in the short term but this is absolutely horrifying censorship and it is only the start.

All of these structures and checks and balances are theoretical: they have never been stress tested in this manner before. Laws are meaningless if not enforced.

The question is: how much will -unelected- Elon Musk be allowed to dismantle the government in this clearly unconstitutional manner.

This has been the darkest week for science in recent memory. Horrifying to imagine what else is to come.

I like to think there are others out there on their own improbable journeys. These are the students who perhaps stand to gain the most if successful, and we do them a great disservice when we give up on them before giving them a chance because they deviate from an increasingly-precise template

end/

The dilemma is not just who fails when we select candidates purely on numbers, but who are we excluding that would have succeeded? What do these metrics inadequately model about what makes someone a successful research scientist?

8/

Which is to say, academia is not alone in this problem of using indicators of past performance to predict future success - particularly when the past performance insufficiently captures the physical, mental, and emotional tolls to be endured in future training. 7/

How to select SEAL Candidates who will be successful in training has been a long-standing problem. It turns out that physical fitness scores are not sufficiently predictive: grit is also a critical component - but difficult to measure. 6/

It reminds me of a similar problem the military faces: How do you assess and select Navy SEALS?

BUD/S, the school for SEAL Candidates, has a physical fitness test that represents a high barrier to entry - and yet up to 90% of those that enter fail to become SEALS.

5/

What do you do when the answer isn't in the back of the book? When nobody knows the answer and you only have yourself to rely on?

Students used to achieving high scores and riding the coattails of impressive team projects may flounder in conditions of relentless failure and uncertainty. 4/

I felt intimidated and hopelessly out of my depth the first week among students who were more prepared or from better schools

And yet... All those smart students who knew all the answers week 1? Most of them didn't make it to the end. What I lacked in knowledge I made up for in mental toughness 3/

I came from a small, unranked public school with poor test scores, a poor math background, and one co-author manuscript in preparation. If you lined me up in class rank I was probably near the bottom. 2/