FSPool: Learning Set Representations with Featurewise Sort Pooling

Abstract

Traditional set prediction models can struggle with simple datasets due to an issue we call the responsibility problem. We introduce a pooling method for sets of feature vectors based on sorting features across elements of the set. This can be used to construct a permutation-equivariant auto-encoder that avoids this responsibility problem. On a toy dataset of polygons and a set version of MNIST, we show that such an auto-encoder produces considerably better reconstructions and representations. Replacing the pooling function in existing set encoders with FSPool improves accuracy and convergence speed on a variety of datasets.

Published at International Conference on Learning Representations (ICLR) 2020

References in corpus (10)

• Learning Representations and Generative Models for 3D Point Clouds
• Gather-Excite: Exploiting Feature Context in Convolutional Neural Networks
• GraphRNN: Generating Realistic Graphs with Deep Auto-regressive Models
• Learning Latent Permutations with Gumbel-Sinkhorn Networks
• Sorting out Lipschitz function approximation
• Deep Set Prediction Networks
• Stochastic Optimization of Sorting Networks via Continuous Relaxations
• Deep Perm-Set Net: Learn to predict sets with unknown permutation and cardinality using deep neural networks
• Loss Functions for Multiset Prediction
• Learning Representations of Sets through Optimized Permutations

Cited by in corpus (18)

• Object-Centric Learning with Slot Attention
• Deep Set Prediction Networks
• The LHC Olympics 2020: A Community Challenge for Anomaly Detection in High Energy Physics
• Decomposing 3D Scenes into Objects via Unsupervised Volume Segmentation
• Conditional Set Generation with Transformers
• Learning the Best Pooling Strategy for Visual Semantic Embedding
• Set-to-Sequence Methods in Machine Learning: a Review
• Studying The Effect of MIL Pooling Filters on MIL Tasks
• Set Representation Learning with Generalized Sliced-Wasserstein Embeddings
• On Adversarial Robustness of 3D Point Cloud Classification under Adaptive Attacks
• Better Set Representations For Relational Reasoning
• Neural Auction: End-to-End Learning of Auction Mechanisms for E-Commerce Advertising
• Multiset-Equivariant Set Prediction with Approximate Implicit Differentiation
• Particle Graph Autoencoders and Differentiable, Learned Energy Mover's Distance
• USCO-Solver: Solving Undetermined Stochastic Combinatorial Optimization Problems
• Top-N: Equivariant set and graph generation without exchangeability
• Set Prediction without Imposing Structure as Conditional Density Estimation
• multiPRover: Generating Multiple Proofs for Improved Interpretability in Rule Reasoning