Augmenting the User-Item Graph with Textual Similarity Models
arXiv:2109.09358
Abstract
This paper introduces a simple and effective form of data augmentation for recommender systems. A paraphrase similarity model is applied to widely available textual data, such as reviews and product descriptions, yielding new semantic relations that are added to the user-item graph. This increases the density of the graph without needing further labeled data. The data augmentation is evaluated on a variety of recommendation algorithms, using Euclidean, hyperbolic, and complex spaces, and over three categories of Amazon product reviews with differing characteristics. Results show that the data augmentation technique provides significant improvements to all types of models, with the most pronounced gains for knowledge graph-based recommenders, particularly in cold-start settings, leading to state-of-the-art performance.
12 pages, 2 figures
References in corpus (11)
- An Overview of Multi-Task Learning in Deep Neural Networks
- KGAT: Knowledge Graph Attention Network for Recommendation
- Convolutional Neural Networks for Sentence Classification
- RotatE: Knowledge Graph Embedding by Relational Rotation in Complex Space
- Unifying Knowledge Graph Learning and Recommendation: Towards a Better Understanding of User Preferences
- Translation-based Recommendation
- Inferring Networks of Substitutable and Complementary Products
- How Useful are Reviews for Recommendation? A Critical Review and Potential Improvements
- Latent geometry of bipartite networks
- Ricci Curvature of the Internet Topology
- Symmetric Spaces for Graph Embeddings: A Finsler-Riemannian Approach