Our paper “Tractable Representation Learning with Probabilistic Circuits” has been accepted to TMLR! This work develops methods for tractable representation learning using probabilistic circuits, enabling efficient and interpretable deep learning.