I'd like to discuss (get you interested into) a problem that unifies and improves the theory from two seemingly unrelated STOC papers (so other people beside me think this is interesting). The theory can be used for various things like proving polynomial smoothed complexity for Integer linear programming, bounding the number of Pareto optimal solutions for multicriteria optimization problems, or having a great time while thinking about a solution for the problem. The problem itself is easy to state (but not easy enough for this abstract:), no special knowledge is required.
This will be an informal talk (so casual clothing is OK :)