SetVAE: Learning Hierarchical Composition for Generative Modeling of Set-Structured Data

CVPR 2021  ·  Jinwoo Kim, Jaehoon Yoo, Juho Lee, Seunghoon Hong ·

Generative modeling of set-structured data, such as point clouds, requires reasoning over local and global structures at various scales. However, adopting multi-scale frameworks for ordinary sequential data to a set-structured data is nontrivial as it should be invariant to the permutation of its elements. In this paper, we propose SetVAE, a hierarchical variational autoencoder for sets. Motivated by recent progress in set encoding, we build SetVAE upon attentive modules that first partition the set and project the partition back to the original cardinality. Exploiting this module, our hierarchical VAE learns latent variables at multiple scales, capturing coarse-to-fine dependency of the set elements while achieving permutation invariance. We evaluate our model on point cloud generation task and achieve competitive performance to the prior arts with substantially smaller model capacity. We qualitatively demonstrate that our model generalizes to unseen set sizes and learns interesting subset relations without supervision. Our implementation is available at

PDF Abstract CVPR 2021 PDF CVPR 2021 Abstract
Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Point Cloud Generation ShapeNet Airplane SetVAE MMD-CD 0.199 # 1
1-NNA-CD 75.31 # 4
Point Cloud Generation ShapeNet Car SetVAE MMD-CD 0.88 # 1
1-NNA-CD 59.66 # 3
Point Cloud Generation ShapeNet Chair SetVAE MMD-CD 2.55 # 2
1-NNA-CD 58.76 # 3