Worthwhile reads from the past week:
Name some problems where frequentists have gotten better results than Bayesians.
Probabilities generate frequencies not the other way round.
That’s not so much a Thought for the Day as it is one of life’s guiding principles.
(“What would John Tukey do?”)
I’m attending the Computer Vision and Pattern Recognition (CVPR) conference this week. I’ve read a lot of papers from past proceedings but this is the first time I’ve attended. (A few interesting talks and posters but so far nothing particularly applicable to work. Still a few days to go though.) Anyhow, best quote so far from the meeting:
You might think that you can move in any direction in the tomato space.
Ah, but you can’t! In all seriousness, the speaker was talking about a mathematical representation of tomatoes from unripe to ripe to sliced or diced. More generally, his topic was representation of objects undergoing transformations. It was an interesting talk.