Dual Gauss-Newton Directions for Deep Learning.
Vincent RouletMathieu BlondelPublished in: CoRR (2023)
Keyphrases
- deep learning
- gauss newton
- convergence analysis
- levenberg marquardt
- unsupervised learning
- active appearance models
- machine learning
- newton method
- weakly supervised
- mental models
- gaussian process
- least squares
- primal dual
- model selection
- global convergence
- semi supervised
- artificial neural networks
- bayesian networks
- computer vision