zlacker

Are there theoretical models that use real time weights? Every intro to deep learning focuses on stochastic gradient descent for neural network weights; as a layperson I'm curious about what online algorithms would be like instead.