A variational Bayes approach to debiased inference for low-dimensional parameters in high-dimensional linear regression.
Ismaël CastilloAlice L'HuillierKolyan RayLuke TravisPublished in: CoRR (2024)
Keyphrases
- linear regression
- variational bayes
- low dimensional
- high dimensional
- hyperparameters
- parameter space
- nonlinear regression
- bayesian inference
- least squares
- latent variables
- gaussian mixture model
- dimensionality reduction
- model selection
- maximum likelihood
- expectation maximization
- posterior distribution
- high dimensional data
- closed form
- em algorithm
- bayesian learning
- cross validation
- free energy
- prior information
- latent dirichlet allocation
- parameter estimation
- exponential family
- gaussian process
- input space
- support vector
- dimension reduction
- maximum a posteriori
- bayesian framework
- feature space
- log likelihood
- sample size
- data points
- probability density function
- random sampling
- principal component analysis
- incomplete data
- noise level
- mixture model
- markov chain monte carlo
- lower bound
- computer vision
- input data
- parameter settings
- probability distribution
- approximate inference
- probabilistic model
- incremental learning
- machine learning
- decision trees