Title: Bayesian variable selection in high dimensional problems without assumptions on prior model probabilities

Abstract: We consider the problem of variable selection in linear models when $p$, the number of potential regressors, may exceed (and perhaps substantially) the sample size $n$ (which is possibly small).