A robust autoregressive gaussian process motion model using l1-norm based low-rank kernel matrix approximation.
Eunwoo KimSungjoon ChoiSonghwai OhPublished in: IROS (2014)
Keyphrases
- low rank
- gaussian process
- motion model
- autoregressive
- low rank matrix
- semi supervised
- kernel matrix
- norm minimization
- convex optimization
- matrix completion
- matrix factorization
- missing data
- model selection
- motion estimation
- linear combination
- optical flow
- particle filter
- regression model
- approximate inference
- singular value decomposition
- image sequences
- random fields
- bayesian framework
- low rank approximation
- latent variables
- hyperparameters
- high dimensional data
- support vector
- kernel methods
- pairwise
- prior knowledge
- machine learning
- feature space
- collaborative filtering
- active learning