If you watched Muthu Muthukrishnan's presentation at the Berkeley Streaming meeting entitled Modern Algorithmic Tools for Analyzing Data Streams you know that Count-Min Sketch can be applied to a variety of situations. We are accustomed to see those in Compressive Sensing as they are connected to some sparse measurement matrices. I am sure something it can also fir within the GraphLab framework. But ....
....here is a new and very interesting one: Particle Sketches, A Way to Estimate a Very Large Number of Bayesian Models in Sub-linear Space and Time.
Now the question is: could this be implemented directly into an 'inexact' chip. ?
Thanks Arthur for the heads-up!
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.