- The utility of randomness: returning from a machine learning conference in Italy where I had a pretty random (if you do not believe in fate) conversation with a stranger, where suddenly more people stood around and joined in. It basically started with me commenting on one poster "Isn't it amazing how useful randomness is, given there is no pattern/structure in a random sequence, and it costs me nothing to make up?" What I meant is how can something as arbitrary still be useful. Suddenly we started collecting examples: randomized algorthms like the random walker model behind PageRank, Random Forests in machine learning, random numbers for perfect encryption (one time pad), Pentti Kanerva's "Hyperdimensional Computing: An Introduction to Computing in Distributed Representation with High-Dimensional Random Vectors" (thanks to an anonymous Finnish bystander, for I could not recall Kanerva's name at the time) etc. Today, I found a book that studies how random people's travel decisions are: Ennio Cascetta (2009) Random Utility Theory, Heidelberg: Springer.

- hypergraphs: graphs have been very useful representations for so many things in work and life, but they are only for expressing dyadic relations between nodes, whereas their generalizations, hypergraphs can express n-ary relations between nodes.

- how to bring people back to a culture of books & libraries: I collect books and I read a lot. As I teach/lecture also, I often ask my students about their reading habits, and find it shocking that many have never been to a library, and most do not read books. I would like to contribute to changing that (seems harder than finding out whether P = NP?).