Federated Knowledge Graph Completion via Latent Embedding Sharing and Tensor Factorization
0
Citation
0
Reference
10
Related Paper
Abstract:
Knowledge graphs (KGs), which consist of triples, are inherently incomplete and always require completion procedure to predict missing triples. In real-world scenarios, KGs are distributed across clients, complicating completion tasks due to privacy restrictions. Many frameworks have been proposed to address the issue of federated knowledge graph completion. However, the existing frameworks, including FedE, FedR, and FEKG, have certain limitations. = FedE poses a risk of information leakage, FedR's optimization efficacy diminishes when there is minimal overlap among relations, and FKGE suffers from computational costs and mode collapse issues. To address these issues, we propose a novel method, i.e., Federated Latent Embedding Sharing Tensor factorization (FLEST), which is a novel approach using federated tensor factorization for KG completion. FLEST decompose the embedding matrix and enables sharing of latent dictionary embeddings to lower privacy risks. Empirical results demonstrate FLEST's effectiveness and efficiency, offering a balanced solution between performance and privacy. FLEST expands the application of federated tensor factorization in KG completion tasks.Keywords:
Graph Embedding
Knowledge graph
Node embedding learns a low-dimensional representation for each node in the graph. Recent progress on node embedding shows that proximity matrix factorization methods gain superb performance and scale to large graphs with millions of nodes. Existing approaches first define a proximity matrix and then learn the embeddings that fit the proximity by matrix factorization. Most existing matrix factorization methods adopt the same proximity for different tasks, while it is observed that different tasks and datasets may require different proximity, limiting their representation power.
Graph Embedding
Representation
Matrix (chemical analysis)
Cite
Citations (25)
In this thesis, we address the spectral and J-spectral factorization problems. We first review the multivariate spectral factorization method devised by Youla in his celebrated paper (Youla, 1961) focusing, in particular, on some of its remarkable features. Then, in the spirit of Youla’s work, we present a technique which provide a solution to the multivariate spectral factorization problem in discrete-time. Finally, a J-spectral extension of the proposed factorization approach is discussed
Spectral theorem
Matrix (chemical analysis)
Cite
Citations (1)
We study processes with unstable particles in intermediate time-like states. It is shown that the amplitudes squared of such processes factor exactly in the framework of the model of unstable particles with continuous masses. Decay widths and cross sections can then be represented in a universal factorized form for an arbitrary set of interacting particles. This exact factorization is caused by specific structure of propagators in the model. We formulate the factorization method and perform a phenomenological analysis of the factorization effects. The factorization method considerably simplifies calculations while leading to compact and reasonable results.
Propagator
Dixon's factorization method
Cite
Citations (1)
Weierstrass factorization theorem
Cite
Citations (0)
Recent works have discussed the violation of factorization and universality in hadronic hard scattering processes aimed at measurements of T-odd distributions. We use simple arguments to show that it is possible to restore an approximate factorization involving T-odd contributions if the factorization breaking interactions present a frequency spectrum dominated by by a narrow and regular peak whose maximum value corresponds to a respected factorization.
Cite
Citations (0)
A Schur-based factorization procedure is described for impedance matrices transformed into a localizing basis. A recently reported factorization strategy relying on a sparse QR factorization of the transformed impedance matrix was shown to exhibit improved efficiency relative to standard direct factorizations. It is shown here that a Schur factorization provides additional efficiencies.
QR decomposition
Basis (linear algebra)
LU decomposition
Matrix (chemical analysis)
Impedance parameters
Cite
Citations (19)
We give an overview of the current status of perturbative QCD factorization theorems in processes that involve transverse momentum dependent (TMD) parton distribution functions (PDFs) and fragmentation functions (FF). We enumerate those cases where TMD-factorization is well-established, and mention cases where it is likely to fail. We discuss recent progress in the implementation of specific TMD-factorization calculations, including the implementation of evolution. We also give examples of hard part calculations. We end by discussing future strategies for the implementation of TMD-factorization in phenomenological applications.
Weierstrass factorization theorem
Cite
Citations (4)
Cite
Citations (14)
We analyse two consequences of the relationship between collinear factorization and $k_t$-factorization. Firstly we show that the $k_t$-factorization gives a fundamental justification for the choice of $Q^2$ (see the text for a precise definition of this variable) done in the collinear factorization. Secondly, we show that in the collinear factorization there is an uncertainty on this choice which will not be reduced by higher orders. This uncertainty is absent within the $k_t$-factorization formalism.
Formalism (music)
Cite
Citations (0)
Translation-based knowledge graph embedding has been one of the most important branches for knowledge representation learning since TransE came out. Although many translation-based approaches have achieved some progress in recent years, the performance was still unsatisfactory. This paper proposes a novel knowledge graph embedding method named TripleRE with two versions. The first version of TripleRE creatively divide the relationship vector into three parts. The second version takes advantage of the concept of residual and achieves better performance. In addition, attempts on using NodePiece to encode entities achieved promising results in reducing the parametric size, and solved the problems of scalability. Experiments show that our approach achieved state-of-the-art performance on the large-scale knowledge graph dataset, and competitive performance on other datasets.
Knowledge graph
ENCODE
Graph Embedding
Feature Learning
Cite
Citations (8)