boxfire

Any problem where the data is imprecise or the model itself is very approximate. Machine learning type problems, machine vision, as the paper shows interferometry, lossy image processing. I can think of a few more data processing problems, but I think you might get the point. The use of a precision processor even in single precision when your input data is already 5%+ noise is a huge waste. This has immense application.

Having dæmon instead of daemon in the article makes my blood boil a little, but I managed to push through. The ligature has no place in writing about computer daemons. It might apply if we were having a talk about medieval literature regarding demons.

I find the Kalman filter explanation in terms of the Chokesky decomposition by R. Eubank to be excellent. It has a prerequisite of knowing linear regression (the linear algebra actually know what you are doing type), but made the inference of what the Kalman filter does very clear. This is definitely not a 10 minute blog post explanation, but understanding how it functions on a deep level is worth it.

http://read.pudn.com/downloads158/ebook/707373/A_Kalman_Filt...

I don't know why people jump through hoops to start out with a multidimensional Kalman filter, and even sometimes dive deep into complex topics. The subject seems to suffer from the same terminology and confusion gap as pre-Kolmogorov statistics.

I would LOVE to see some form of PANDA integration with a tool like this. https://github.com/moyix/panda Having this kind of analysis on a PANDA replay / snapshot would be great.

The combination of the two would be a perfect reverse engineering toolbox. This is great on its own and to use side by side though.

Originally in my terminal I was experimenting to find a cutoff for the point where the numbers become inaccurate. All of those formulas were in terms of digits and had a -1 on the end.

> echo 'scale=500-1;10/(10^20-10^10-1)' | BC_LINE_LENGTH=12 bc

will give you the ten digit ones, with the last digit rounded up on the very end.

I havent quite got an exact formula for scale before the rounding is wrong, e.g. for 100 digits:

> echo 'scale=50000-2000-1;10/(10^200-10^100-1)' | BC_LINE_LENGTH=102 bc