Space shuttle model: A physics inspired method for learning quantizable deep representations.
Shicong LiuHongtao LuPublished in: ICME (2017)
Keyphrases
- prior knowledge
- learning scheme
- mathematical model
- statistical model
- classification method
- probabilistic model
- linear model
- prediction model
- evaluation method
- parameter estimation
- learned models
- learning mechanism
- optimization method
- theoretical analysis
- modeling method
- cost function
- automatically learned
- boosted decision trees
- privileged information
- neural nets
- high precision
- machine learning
- input data
- high accuracy
- supervised learning
- objective function
- similarity measure
- bayesian framework
- bp neural network
- em algorithm
- bayesian methods
- probability distribution
- reinforcement learning
- neural network
- space shuttle
- deep belief networks
- autoregressive
- generative model
- unsupervised learning