Transformers for Complex Query Answering over Knowledge Hypergraphs
Published in Arxiv, 2025
The study introduces two new CQA datasets, JF17k-HCQA and M-FB15k-HCQA, to better model complex real-world data using knowledge hypergraphs. A novel two-stage transformer model, the Logical Knowledge Hypergraph Transformer (LKHGT), is proposed for answering complex logical queries, demonstrating state-of-the-art performance on CQA tasks and generalizing well to new query types.