Optimizing Nondecomposable Loss Functions in Structured Prediction.
Mani RanjbarTian LanYang WangStephen N. RobinovitchZe-Nian LiGreg MoriPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2013)
Keyphrases
- loss function
- structured prediction
- maximum margin
- support vector
- conditional random fields
- pairwise
- latent variables
- efficient learning
- markov networks
- approximate inference
- convex optimization
- learning models
- machine learning
- graphical models
- belief networks
- hyperplane
- learning algorithm
- maximum likelihood
- support vector machine