Published: Sep 17, 2013 by Gavin
Greetings fine people of the internet and welcome to my little corner of it.
I intend to use this space to discuss my research and any other science I find interesting, although I expect that it will be interspersed with other topics I feel compelled to discuss.
In recent news, after one year of work and a bunch of iterations the model that I am developing and testing with my postdoc supervisor is finally producing results that match with the established ways of calculating these things.
Blue dots show results from traditional way of doing things and the red line shows the results from our model.
As you can see in the unlabeled graph above, the model and traditional way of doing things produce curves with the same slope, but offset y-intercepts (we are currently working on fixing that). Note that I am being deliberately vague here because this work is currently unpublished and I can’t say too much until after it is.
Over 6000 hours of CPU time (a bit more than 250 days) went into acquiring all the data that went into the blue curve. The red curve, on the other hand, simply used 9 hours of CPU time. Even if we can’t get the y-intercept issue sorted out, there is still usefulness in being able to get some predictive data in 0.15% of the time.