Canonical Tensor Decomposition for Knowledge Base Completion
arXiv:1806.07297
Abstract
The problem of Knowledge Base Completion can be framed as a 3rd-order binary tensor completion problem. In this light, the Canonical Tensor Decomposition (CP) (Hitchcock, 1927) seems like a natural solution; however, current implementations of CP on standard Knowledge Base Completion benchmarks are lagging behind their competitors. In this work, we attempt to understand the limits of CP for knowledge base completion. First, we motivate and test a novel regularizer, based on tensor nuclear -norms. Then, we present a reformulation of the problem that makes it invariant to arbitrary choices in the inclusion of predicates or their reciprocals in the dataset. These two methods combined allow us to beat the current state of the art on several datasets with a CP decomposition, and obtain even better results using the more advanced ComplEx model.
References in corpus (5)
- A Review of Relational Machine Learning for Knowledge Graphs
- Restricted strong convexity and weighted matrix completion: Optimal bounds with noise
- Estimation of low-rank tensors via convex optimization
- Learning with the Weighted Trace-norm under Arbitrary Sampling Distributions
- Fast Linear Model for Knowledge Graph Embeddings
Cited by in corpus (13)
- Role-Aware Modeling for N-ary Relational Knowledge Bases
- Knowledge Hypergraph Embedding Meets Relational Algebra
- DyERNIE: Dynamic Evolution of Riemannian Manifold Embeddings for Temporal Knowledge Graph Completion
- MTC: Multiresolution Tensor Completion from Partial and Coarse Observations
- Using Local Knowledge Graph Construction to Scale Seq2Seq Models to Multi-Document Inputs
- Binarized Knowledge Graph Embeddings
- Decompressing Knowledge Graph Representations for Link Prediction
- Software Engineering Event Modeling using Relative Time in Temporal Knowledge Graphs
- Augmenting the User-Item Graph with Textual Similarity Models
- Combining Rules and Embeddings via Neuro-Symbolic AI for Knowledge Base Completion
- CNN-based Dual-Chain Models for Knowledge Graph Learning
- Multi-version Tensor Completion for Time-delayed Spatio-temporal Data
- Path Ranking with Attention to Type Hierarchies