The field of watermarking is one of increasing importance thanks to its application to copyright protection. Unlike coding theory and cryptography, it lacks a solid mathematical foundation. In earlier work, the author proposed a definition of watermarking scheme that includes several schemes currently in use. These schemes come with competing parameters, analogous to rate and minimum distance for codes, that measure how good they are. In this talk, I shall review the current state of the subject, give explicit examples of good schemes, and outline theoretical limits for the parameters, analogous to bounds from coding theory.

This talk will be accessible to a general audience.

In this talk, I will present attacks on security processors that exploit weakensses in their API. The basic idea is that by presenting valid commands to the processor in an unexpected sequence, one can break the security policy envisioned by the designer. Designing APIs that resist such attacks becomes increasingly difficult as the functionality of the processor increases. Preventing such attacks is a challenge facing the formal methods community.

This is a summary of research done by Mike Bond (http://www.cl.cam.ac.uk/~mkb23/) at Cambridge University, UK.