today

I listened to A. Flecher's talk about her work on "Generalized approximation of message passing" (GAMP) this afternoon. The title is attractive, but the talk is not so enjoyable. I think not only for me, but also for the prominent reserachers in redwood center. She is smart, and connnected compressed sensing, sparsity signal recovery and group sparsity method together, meanwhile the idea to connect the sparsity formulation to graphical model and then applied the message passing is clever.

 

However, she is very protective and speak too fast, went over the slides really fast with an high speed of speaking. To my observation, Bruno is trying to understand more, but we don't know he got it or not. It's probabilty a good idea to have a look at their NIPS11 paper after it comes out.

 

I went to the bank with Sealeen today, so ddin't finish reading our papers. We could be more focused.