Connes Kreimer Marcolli
Its about turning knots into simple Feynman diagrams into Multiple Zeta Values. These MZVs satisfy all sorts of crazy relations, which the mathematicans have been studying like crazy. But really they're quite simple. They act on a set of k ordinals (yes, that's right, you should be thinking 1-ordinals) and are characterised by two numbers, namely the weight n, which is the sum of these, and k itself, the so called depth. Of course these naturally show up as special integrals of something called Mixed Tate Motives (don't even ask), so we know that the weight n is the same n of M(0,n+3). Goodness, me. The Yang-Mills problem and the Riemann hypothesis seem to be related. Well, well.
The real question, however, is how to go beyond scalars to other entities in QFT. Any guesses?