Stochastic Downsampling for Cost-Adjustable Inference and Improved Regularization in Convolutional Networks.
Jason KuenXiangfei KongZhe LinGang WangJianxiong YinSimon SeeYap-Peng TanPublished in: CoRR (2018)
Keyphrases
- social networks
- sum product
- network structure
- flow network
- probabilistic inference
- bayesian networks
- probabilistic networks
- bayesian inference
- belief networks
- message passing
- total cost
- expected cost
- small world
- exact inference
- prior information
- inference process
- conditional probability distributions
- complex networks
- tree grammars
- linear gaussian
- monte carlo