Exchangeable deep neural networks for set-to-set matching and learning

Abstract

Matching two different sets of items, called heterogeneous set-to-set matching problem, has recently received attention as a promising problem. The difficulties are to extract features to match a correct pair of different sets and also preserve two types of exchangeability required for set-to-set matching: the pair of sets, as well as the items in each set, should be exchangeable. In this study, we propose a novel deep learning architecture to address the abovementioned difficulties and also an efficient training framework for set-to-set matching. We evaluate the methods through experiments based on two industrial applications: fashion set recommendation and group re-identification. In these experiments, we show that the proposed method provides significant improvements and results compared with the state-of-the-art methods, thereby validating our architecture for the heterogeneous set matching problem.

References in corpus (23)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications
• In Defense of the Triplet Loss for Person Re-Identification
• PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
• Random Erasing Data Augmentation
• Attention-based Deep Multiple Instance Learning
• Stacked Hourglass Networks for Human Pose Estimation
• MMD GAN: Towards Deeper Understanding of Moment Matching Network
• Learning Fashion Compatibility with Bidirectional LSTMs
• Graph Matching Networks for Learning the Similarity of Graph Structured Objects
• Performance Measures and a Data Set for Multi-Target, Multi-Camera Tracking
• Order Matters: Sequence to sequence for sets
• Learning Fine-grained Image Similarity with Deep Ranking
• Group Re-Identification with Multi-grained Matching and Integration
• Deep Mutual Learning
• Building Deep Networks on Grassmann Manifolds
• Universal approximations of permutation invariant/equivariant functions by deep neural networks
• Outfit Generation and Style Extraction via Bidirectional LSTM and Autoencoder
• Convolutional Set Matching for Graph Similarity
• Learning Type-Aware Embeddings for Fashion Compatibility
• FashionNet: Personalized Outfit Recommendation with Deep Neural Network
• On Universal Equivariant Set Networks
• Recommending Outfits from Personal Closet

Cited by in corpus (1)

• The CAT SET on the MAT: Cross Attention for Set Matching in Bipartite Hypergraphs