Expanding End-to-End Question Answering on Differentiable Knowledge Graphs with Intersection
This addresses the limitation of previous models that only handled single-entity questions, improving accuracy for question answering systems on knowledge graphs, though it is incremental as it builds on existing differentiable knowledge graph frameworks.
The paper tackles the problem of handling multiple-entity questions in end-to-end question answering on differentiable knowledge graphs by introducing an intersection operation, resulting in performance improvements from 69.6% to 73.3% Hits@1 on WebQuestionsSP and from 39.8% to 48.7% Hits@1 on ComplexWebQuestions.
End-to-end question answering using a differentiable knowledge graph is a promising technique that requires only weak supervision, produces interpretable results, and is fully differentiable. Previous implementations of this technique (Cohen et al., 2020) have focused on single-entity questions using a relation following operation. In this paper, we propose a model that explicitly handles multiple-entity questions by implementing a new intersection operation, which identifies the shared elements between two sets of entities. We find that introducing intersection improves performance over a baseline model on two datasets, WebQuestionsSP (69.6% to 73.3% Hits@1) and ComplexWebQuestions (39.8% to 48.7% Hits@1), and in particular, improves performance on questions with multiple entities by over 14% on WebQuestionsSP and by 19% on ComplexWebQuestions.