Talks

I heard a few talks the past few days:

• Adam Smith spoke about nifty ideas of differential privacy and described the angst of formulating a workable notion of privacy preserving data analysis.
• Asaf Shapira spoke about property testing, starting with his general results for monotone graph properties for dense graph, and successively nailing down general results for hereditary properties for bounded degree graphs, minor-closed properties of sparse graphs, and testing graph expansion.
  
• Vahab Mirrokni spoke about approximation algorithms for submodular function optimization -- monotone, non-monotone and negative -- and a whirl of problems where they are, or can be found, or planted if one is clever.
  

I feel sated, but I must prepare now to give my upcoming talks (31,1).