When I interviewed last week, I went there to interview a professor. He talked to me for half an hour and handed me off to a [cute and smart] grad student in his group. I talked with him for an hour. They were both trying to explain to me what their research was (which is in "theory of filtering and estimation", using Bayesian inference and a little bit of control theory), and it took me forever to get.
I think the grad student understood that, since he reassured me that it had taken him months to get an intuition about this sort of thing, and I was learning it all in one hour.
To make it short, they're trying to make noisy measurements like radar and precision force sensors more accurate by comparing the sensor measurement to a model, and calling the difference the "innovation error", and using that difference to keep updating the prediction until the error approaches zero. The simulations he showed me were pretty amazing, especially when I knew the math behind it had to be ridiculously complex.
It's going well, though. It took me two days to go through all the material. I did, after all, have an hour and a half of recordings to go through (some of it more than once, and I'm going to have to go back and get more quotes, unfortunately) as well as 7 pages of notes and scribbled diagrams that they had both drawn for me as a way of helping me understand. I've since drawn my own, paraphrased notes on the papers.
The most exciting thing is that this is really my first serious piece of science writing. This is complicated stuff. Not everyone would be able to understand this like I did, with my physics and math background. And most people wouldn't care anyway. So how do I make them understand? And how do I make them care?
This is what science writing is.
If I do this right, this'll be a very important piece for me.
No comments:
Post a Comment