CS: Around the blogs in 80 hours

On Twitter first:

Guiseppe asks

in your surveys on CS have you ever seen a problem like min_X norm1(A*X-B) + k*nuclearnorm(inv(X)) ?

More info can be found here.

Pierre commented on a comment I added to a retweet:

The reason NB has such a high impact "About 200,000 academic journals are published, average number of readers per article is 5" NB it's 50+

For the blogs here is what we have this week:

Bob Sturm:

• Simulation Bugs?
• Paper of the Day (Po'D): Music Genre Classification this time with Sparse Representations Edition
• Paper of the Day (Po'D): Music Genre Classification Edition, no. 2
• Paper of the Day (Po'D): Music Genre Classification Edition, no. 3
• Paper of the Day (Po'D): Multiscale MFCCs Edition

Djalil Chafai

• Piecewise Deterministic Markov Processes

Image Sensor World:

• SiOnyx Announces Record Detectivity (D*) With its New Light Detectors. Three orders of magnitude improvement over current technology, that ought to be a good chip for compressive sensing purposes (i.e. much less noise)
  

Anand Sarwate

• Explicit constructions of Expanders and LDPC codes

Suresh

• The Shape of Shape Analysis Research, Part III . hey Suresh, your blog is blacklisted it is not compatible with public use
• Bad Research As Spam

Frank Nielsen

• Jensen-Bregman Voronoi diagrams.
• Total Bregman Divergence and its Applications to Shape Retrieval . Too bad there is no link to the actual paper and no way to send comments if you do not read Japanese.

Gonzalo Vazquez Vilar

• Cognitive Radio: half empty or half full?
  

Lianlin Li is back

Tanya Khovanova

• My First Polymath Project on a subject that seems eerily close to the 12 ball problem I featured here before. The math is in trying to figure out the bounds for the numbre of measurements.

Sara

• Machine Learning Videos

Andrew Gelman

• Classics of statistics

Meena

• White Matter Fiber Analysis

James Lee

• The Gödel Prize, TSP, and volume growth
• Random partitions of metric spaces
• The generic chaining
• Majorizing measures
• The need for non-linear mappings