In this project, I’m diving into Variational Autoencoders (VAEs) to explore how hidden, or latent, features might directly relate to movement data.
As a generative model, the VAE enables me to identify which latent dimensions correspond to specific characteristics of kinematic patterns—helping to uncover structure that isn’t directly visible in raw data.
Along the way, I’ll use engaging visualizations to clearly explain the internal logic of the model and walk through my thinking and interpretative process.
If you see potential here and feel even half as excited as I do, keep an eye out for the project updates!
