A geometric interpretation of stochastic gradient descent using diffusion metrics.
Rita FioresiPratik ChaudhariStefano SoattoPublished in: CoRR (2019)
Keyphrases
- geometric interpretation
- stochastic gradient descent
- loss function
- least squares
- step size
- matrix factorization
- random forests
- weight vector
- regularization parameter
- importance sampling
- support vector machine
- online algorithms
- multiple kernel learning
- collaborative filtering
- principal component analysis
- learning algorithm
- support vector
- e learning