I left the meterial on hash functions in the notes (pages 13-17) in
case anyone is interested, even though this material wasn't
covered (and, of course, won't be tested on the final exam). I was
planning to do it in the class that I ended up
missing while I was in in Berlin. Instead,
Prof. Selman covered a fascinating connection between probability
and logic instead, showing that although you can't tell easily
whether a formula in what's called k-CNF is satisfiable, you can
find a truth assignment that satisfies "most of it": at least 7/8 of
the clauses in the formula. Although this result doesn't mention probability,
the proof of it uses probabilistic techniques. Here are
Prof. Selman's slides, in powerpoint and
PDF.