You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the 2022-04-26 version of vol 2, sec 9.3.7, there is an example of Gaussian VI applied to a toy 2d logistic regression example (implemented by this JAX script). Modify this example to make it more interesting, and compare posterior of the parametes (not just posterior predictive) to blackJAX.
One possibility is to modify the Africa GDP regression example from sec 9.3.6 by discretizing the response variable (log GDP), but keeping the rest of the model the same.
The example should be solved with full-rank, rank-1, diagonal and HMC.
The text was updated successfully, but these errors were encountered:
murphyk
changed the title
Update "9.3.7 Example: Full-rank vs rank-1 GVI for logistic regression"
Update "Example 9.3.7 (1d logistic regression) using Full-rank vs rank-1 GVI"
Apr 27, 2022
In the 2022-04-26 version of vol 2, sec 9.3.7, there is an example of Gaussian VI applied to a toy 2d logistic regression example (implemented by this JAX script). Modify this example to make it more interesting, and compare posterior of the parametes (not just posterior predictive) to blackJAX.
One possibility is to modify the Africa GDP regression example from sec 9.3.6 by discretizing the response variable (log GDP), but keeping the rest of the model the same.
The example should be solved with full-rank, rank-1, diagonal and HMC.
The text was updated successfully, but these errors were encountered: