TBC
Have a go at (in order of 'importance' maybe?)
- 2.45 verify the Wishart is a conjugate prior for the multivariate Gaussian precision matrix.
- 2.19 show the eigenvector expansion makes sense.
- 2.36 sequential estimation of the variance of a 1d Gaussian.
- 2.5 showing that the beta distribution is correctly normalised.
- 3.1 show that tanh and sigmoids basis functions are equivalent
- 3.4 show that noise on linear model inputs is equivalent to weight decay
- 3.8
- 3.11
- 3.18
- 3.24
TBC
TBC
TBC