Description
MEI introduces the multi-partition embedding interaction technique with block term tensor format to systematically address the efficiency--expressiveness trade-off in knowledge graph embedding. It divides the embedding vector into multiple partitions and learns the local interaction patterns from data instead of using fixed special patterns as in ComplEx or SimplE models. This enables MEI to achieve optimal efficiency--expressiveness trade-off, not just being fully expressive. Previous methods such as TuckER, RESCAL, DistMult, ComplEx, and SimplE are suboptimal restricted special cases of MEI.
Papers Using This Method
Emission-Aware Operation of Electrical Energy Storage Systems2025-06-19Teaching LLMs Music Theory with In-Context Learning and Chain-of-Thought Prompting: Pedagogical Strategies for Machines2025-03-28Knowledge Discovery in Optical Music Recognition: Enhancing Information Retrieval with Instance Segmentation2024-08-27Disentangled Noisy Correspondence Learning2024-08-10Mobile Edge Intelligence for Large Language Models: A Contemporary Survey2024-07-09Major Entity Identification: A Generalizable Alternative to Coreference Resolution2024-06-20MEIM: Multi-partition Embedding Interaction Beyond Block Term Format for Efficient and Expressive Link Prediction2022-09-30Convergence of Policy Gradient for Entropy Regularized MDPs with Neural Network Approximation in the Mean-Field Regime2022-01-18Transformer Embeddings of Irregularly Spaced Events and Their Participants2021-12-31DoReMi: First glance at a universal OMR dataset2021-07-16SGD Distributional Dynamics of Three Layer Neural Networks2020-12-30Multi-Relational Embedding for Knowledge Graph Representation and Analysis2020-09-28Multi-Partition Embedding Interaction with Block Term Format for Knowledge Graph Completion2020-06-29Analyzing Knowledge Graph Embedding Methods from a Multi-Embedding Interaction Perspective2019-03-27