On the diversity of multi-head attention
Jian Li, Xing Wang, Zhaopeng Tu, Michael R. Lyu
Topics & Concepts
Computer scienceSubspace topologyRegularization (linguistics)ExploitArtificial intelligenceRepresentation (politics)Linear subspaceTheoretical computer scienceMachine learningSentenceMachine translationMathematicsPoliticsLawGeometryPolitical scienceComputer securityAdvanced Neural Network ApplicationsTopic ModelingDomain Adaptation and Few-Shot Learning