**CANCELLED**Computing and Mathematical Sciences Colloquium

  Community detection and clustering are crucial components to network and data sciences. Nearly all the data of scientific interest either comes as or can be organized as a network, and clusters are one of the key features to be unveiled. This talk establishes the fundamental limits of community recovery for popular network models called stochastic block models. This allows to set rigorous benchmarks for clustering algorithms, and establishes connections with the fundamental limits of data transmission.