I have finally returned from my vacation — which is the reason why “Pursuits in the Null Space” has been so dormant. In the coming months there will be a flurry of activity here, as I gear up to begin in January a two-year research project titled, “Greedy Sparse Approximation and the Automatic Description of Audio and Music Data.” It won’t be entirely devoted to “greedy” methods of course, or exclusively focused on audio and music data. But audio and music data provides a good real world playground with plenty of stuff difficult to climb (sans injury).
And I have some papers accepted to Asilomar this year:
- B. L. Sturm, M. Christensen, and R. Gribonval, “Cyclic Greedy Algorithms for Recovering Compressively Sampled Sparse Signals”.
- M. Christensen and B. L. Sturm, “A Perceptually Re-Weighted Mixed-Norm Method for Sparse Approximation of Audio Signals”.
That should be a good conference as there appears to be several interesting titles in the program. And this time I think I will rent a surfboard.
Regarding my investigations on the effects of distributions on sparse signal recovery, Phil Schniter has let me know about his work in combining approximate message passing with expectation maximization to learn the signal distribution during its recovery. Here is a poster he and a student recently presented at the Duke Workshop on Sensing and Analysis of High-Dimensional Data. The results look promising.
The ICASSP 2012 deadline is a little more than a month away; and soon I will be teaching a new course for me, Multivariate Statistics and Pattern Recognition.
And for this course, I have just completed some notes for my probability review session titled, Three Fun Games of Probability, and the Big Bang Lottery, or, The Fatal Flaw of Arguing for Intelligent Design from Probability.
This will be fun!