Here is what I found relevant this past week on the blogs and elsewhere:

* Two questions on the Compressive sensing LinkedIn group asking if anyone knew of a formulation for total variation minimization using the AMP (approximate message passing) algorithm?, or Identifying a good Metric for Comparison between various CS Reconstruction algorithms. That group has 1481 members, while the Advanced Matrix Factorization group has 292 members.

* Dustin goes deeper in explaining his recent paper on why Certifying the Restricted Isometry Property is Hard.

* Danny tells us about a synthetic large benchmark, Ensemble methods - a notable book by Elder and Seni, a KDD CUP Update and a KDD CUP 2012 Track 1 using GraphLab..

* Bob writes about:
- Experiments with MOD
- A Concatenative Synthesis Patent Challenge?
- A missing step from the Method of Directions?
- Dictionary learning with the Method of Optimal Directions
- Theorem 10 of On polar polytopes and the recovery of sparse representations
- Computer Music Judges

* Suresh talks about Approximate Bregman near neighbors. (and some interesting results) and Distributed Learning: A new model

* Sergey points to the recent Cerex solvers released by folks at Google, several arxiv find here and here and a collaborative LaTeX editor.

* Greg talks about his Paper on the MIT Build a Radar Course that will be presented at the IEEE Radar Con 2012

* Terry talks about math gamification in Lewis Carroll logic puzzles , New version of algebra game and Gamifying algebra?. In order to get people addicted maybe we ought to consider how this gamification of the learning process can be encapsulated into some NP-hard form.

* Finally, John talks about ICML: Behind the Scenes

Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.

## No comments:

Post a Comment