- First, a different topic. Burt Ovrut has convinced me that while a slope-stable bundle in the hidden sector of their models has not been explicitly constructed, doing so is not necessary. It is pretty clear that one can always wrap fivebranes and anti-fivebranes on the appropriate cycles to cancel all anomalies and solve all conditions of string theory - and even show that the moduli of such fivebranes are stabilized. These moduli are in fact stabilized already in the absence of anti-branes, see here. SUSY will generically be broken and the cosmological constant will get a boost to become positive, in a somewhat KKLT-like way. Of course, no anthropic principle is assumed to work here. If you evaluate this data, it is fair to say that there exists a consistent background with properties studied by Ovrut et al. and their approach to focus on the visible bundle and study it separately seems justified.
- two-dimensional RG flow and closed string tachyon dynamics
based on his work with Dan Freedman and Matt Headrick. One of the general questions he wanted to answer is "What is the configuration space of [perturbative] string theory?"
The space of "on-shell" solutions is nothing else than the space of conformal field theories. What about the "off-shell" spacetime configurations in string theory? Albion started with the assumption that the off-shell configurations are all two-dimensional field theories, not necessarily conformal ones.
This led him to study the relations between RG and GR. He included an extra spacetime dimension - assuming that it has a time-like signature - and allowed all fields to depend on this new dimension. This brought him into an interesting spacetime (or just "time") effective action in which the beta function for the dilaton plays the role of the potential energy.
Whoever is interested in the details should click the only link in this text above or see Robert Helling's excited description.