Sequential Monte Carlo Methods to Train Neural Network Models
We discuss a novel strategy for training neural networks using sequential Monte Carlo algorithms and propose a new hybrid gradient descent / sampling importance resampling algorithm (HySIR). In terms of computational time and accuracy, the hybrid SIR is a clear improvement over conventional sequenti...
Autors principals: | Freitas, D, Nando, Niranjan, M, Gee, A, Doucet, A |
---|---|
Format: | Journal article |
Publicat: |
2000
|
Ítems similars
-
Sequential monte carlo methods To train neural network models
per: , d, et al.
Publicat: (2000) -
Sequential Monte Carlo methods for diffusion processes
per: Jasra, A, et al.
Publicat: (2009) -
Sequential Monte Carlo samplers
per: Del Moral, P, et al.
Publicat: (2006) -
Sequential Monte Carlo for model selection and estimation of neural networks
per: Andrieu, C, et al.
Publicat: (2000) -
Controlled sequential Monte Carlo
per: Heng, J, et al.
Publicat: (2020)