Delves into the fundamental limits of gradient-based learning on neural networks, covering topics such as binomial theorem, exponential series, and moment-generating functions.
Covers the definition of multivariate Gaussian distribution and its properties, including moment generating function and linear combinations of variables.
