NIPS
2001
13 years 6 months ago
2001
In this paper we introduce a new sparseness inducing prior which does not involve any (hyper)parameters that need to be adjusted or estimated. Although other applications are poss...