An Exploration of Learnt Representations of W Jets
This work addresses the need for interpretable machine learning representations in high-energy physics, offering incremental improvements in understanding jet data manifolds.
The authors tackled the problem of learning interpretable representations of boosted W jets in collider physics by training a Variational Autoencoder with an Earth Movers Distance-based reconstruction error, resulting in a latent space with semantically meaningful and hierarchically organized directions that provide insight into scale-dependent structure and information complexity.
I present a Variational Autoencoder (VAE) trained on collider physics data (specifically boosted $W$ jets), with reconstruction error given by an approximation to the Earth Movers Distance (EMD) between input and output jets. This VAE learns a concrete representation of the data manifold, with semantically meaningful and interpretable latent space directions which are hierarchically organized in terms of their relation to physical EMD scales in the underlying physical generative process. The variation of the latent space structure with a resolution hyperparameter provides insight into scale dependent structure of the dataset and its information complexity. I introduce two measures of the dimensionality of the learnt representation that are calculated from this scaling.