### Posts in the past four weeks

**Tuesday**

Jan 20 2015

04:41 UTC

... (Ness, Rix, and my **data**-driven model for stellar spectra), which is doing a great job of labeling APOGEE spectra. It is a new talk, so I started from scratch. Big question: Will the projector be 4:3 or 16:9? Of course I can switch my slides with one line of LaTeX.

Posted by Hogg's Research

**Friday**

Jan 16 2015

04:59 UTC

... inexpensively with **data** and simulations in hand (that is, if we can convince some of the BOSS people to help out). Second, we discussed expanding Fadely's work on modeling the SDSS imaging catalo

Posted by Hogg's Research

**Thursday**

Jan 15 2015

04:59 UTC

... gave a secont talk today at Computing the Universe, this time on nuisance parameter models. I continued my discussion of Gaussian processes and then made the following points: You must fit your nuisance simultaneously with your parameters of interest and you must marginalize them out; you don't want to estimate anything you don't care about. Don't fit and subtract or you will be subject to over- or under-fitting. The most conservative thing you can do is implement a model with

Posted by Hogg's Research

**Tuesday**

Jan 13 2015

04:59 UTC

... put my research time today into the discussion section in the nearly finished paper by Ness et al on The Cannon, our project to put parameter and chemical abundance labels on stars without using stellar models. In my current practice, I built a part of the discussion around our assumptions: What assumptions did we make, why did we make them, and what would be the costs and benefits of changing them.

Posted by Hogg's Research

**Thursday**

Jan 08 2015

04:59 UTC

... understanding, in real and fake **data**, how our image model likelihood behaves as a function of position and region size. There were some surprises there, all obvious in retrospect: One was that if there is any gradient or non-stationarity in the background pixels, the model wants to split the background into two (or more) huge boxes. Another was that the success of the m

Posted by Hogg's Research

**Wednesday**

Jan 07 2015

04:59 UTC

Frean and I started coding up our generative image model. Since all that matters is histograms, we only need counts in regions of the image. Frean had a nice trick in which we pre-cumulate what you might call the "indicator" images, one per histogram bin, and then counts in bins become some very simple operations only on the corners of the region. Awesome and fast and very numpythonic. We tested the method by swinging a patch over a SDSS image, and looking at the likelihood of a source-plus-ba

Posted by Hogg's Research

**Saturday**

Jan 03 2015

04:59 UTC

Vacation is over! I spent the train ride from Auckland to Wellington working on the first paper about The Cannon and also my MCMC tutorial. I also thought a bit more about CMB real-space analysis and foregrounds, which led to a breakthrough of sorts:I realized that the ideal model for the CMB foregrounds is not to try to model each band with linear (or non-linear) combinations of other bands; the right idea is that there are a (hopefully small) number of latent variables that control the emiss

Posted by Hogg's Research