Better latent spaces for better autoencoders
Barry M. Dillon
Abstract
Autoencoders as tools behind anomaly searches at the LHC have the structural problem that they only work in one direction, extracting jets with higher complexity but not the other way around. To address this, we derive classifiers from the latent space of (variational) autoencoders, specifically in Gaussian mixture and Dirichlet latent spaces. In particular, the Dirichlet setup solves the problem and improves both the performance and the interpretability of the networks.
Topics & Concepts
InterpretabilityLatent Dirichlet allocationArtificial intelligenceSpace (punctuation)Computer scienceGaussianDirichlet distributionMixture modelAnomaly detectionPattern recognition (psychology)Anomaly (physics)Machine learningMathematicsTopic modelPhysicsCondensed matter physicsQuantum mechanicsMathematical analysisBoundary value problemOperating systemParticle physics theoretical and experimental studiesAlgorithms and Data CompressionBayesian Methods and Mixture Models