Projects, paper reviews, and CV/ML concepts which I like to summarize for my own future reference.
How modern neural nets handle fisheye distortion and metric scale
Modernizing CNNs with Transformer design choices
A different ELBO with the same Jensen trick, then merging it with the VAE
Putting a stochastic latent variable inside a neural network
RL Intro and Imitation Learning
Learning by trial and error
Learning what's good, then doing more of it
Bounding $p(x)$ with the Evidence Lower BOund
Forward and inverse projection, distortion models
