Abstract | ||
---|---|---|
Recent research has seen the proposal of several new inductive principles designed specifically to avoid the problems associ- ated with maximum likelihood learning in models with intractable partition functions. In this paper, we study learning methods for binary restricted Boltzmann machines (RBMs) based on ratio matching and gen- eralized score matching. We compare these new RBM learning methods to a range of ex- isting learning methods including stochastic maximum likelihood, contrastive divergence, and pseudo-likelihood. We perform an ex- tensive empirical evaluation across multiple tasks and data sets. |
Year | Venue | DocType |
---|---|---|
2010 | AISTATS | Journal |
Volume | Citations | PageRank |
9 | 54 | 2.25 |
References | Authors | |
8 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Benjamin Marlin | 1 | 950 | 95.15 |
Kevin Swersky | 2 | 1118 | 52.13 |
Bo Chen | 3 | 1097 | 33.93 |
Nando De Freitas | 4 | 3284 | 273.68 |