Luis

[lrm22005]

I did for my final project 4 different models

One model is using variational gaussian process, with active learning loading data after dimensionality reduction using tsne.
One model is using variational gaussian process with active learning loading the original data.
One model is using Meta learning clustering to Semi supervised Variational gaussian process with active learning loading the original data, labeling additional data.

The complexity of the models is related with the latent spaces defined and the quantity of iterations that it allows. I choose minimum quantity of iterations because I just need to run it to get results on the course, but maybe for the paper we can try more iterations.