pyprobml

pyprobml copied to clipboard

https://github.com/probml/pyprobml/blob/master/notebooks/book2/18/gpKernelPlot.ipynb Change ordering to ``` rbf(0.1), rbf(1), rbf(5) mat12, 32, 52 periodic, cosine, RQ poly1, poly2, noise ``` Put more space between rows Also save as individual figures/files in addition...

murphyk

tweak a tikz figure for sequential prediction

2

tweak https://github.com/probml/pyprobml/blob/master/tikz/sequential_prediction_tikz.tex so it looks like this (adding loss nodes L_t) 

murphyk

Rewrite https://github.com/probml/pyprobml/blob/master/notebooks/book2/29/newsgroupsVisualize.ipynb so it uses [sklearn.datasets.fetch_20newsgroups](https://scikit-learn.org/stable/modules/generated/sklearn.datasets.fetch_20newsgroups.html) instead of downloading the matlab version of the data from probml-data. This will ensure is uses the same daataset as https://github.com/probml/pyprobml/blob/master/notebooks/book2/29/relevance_network_newsgroup_demo.ipynb

murphyk

https://tikz.net/nf-coupling-layer/

murphyk

murphyk

can we replace this with our own code?

murphyk

murphyk

murphyk

tweak the plotting parameters of https://github.com/probml/pyprobml/blob/master/notebooks/book2/13/pf_guided_neural_decoding.ipynb so it doesn't look so ugly

murphyk

implement GP-LVM from scratch

1

Currently fig 28.19 is created by https://github.com/probml/pyprobml/blob/master/notebooks/book2/28/gplvm_mocap.ipynb which uses GPy from Sheffield. Here is the core code ``` model = GPy.models.GPLVM(Y, 2, init="PCA", normalizer=True) model.optimize(optimizer="lbfgs", messages=True, max_f_eval=1e4, max_iters=1e4) ``` It...

murphyk