Online Learning and Optimization from Continuous to Discrete Time
Many discrete algorithms for convex optimization and online learning can be interpreted as a discretization of a continuous-time process. Perhaps the simplest and oldest example is gradient descent, which is the discretization of the ODE…