Strong Baselines for Simple Question Answering over Knowledge Graphs with and without Neural Networks
This work highlights that gains from complex deep learning models in this domain are modest, suggesting unnecessary complexity in prior approaches, which is important for researchers and practitioners in knowledge graph question answering.
The paper tackled simple question answering over knowledge graphs by decomposing it into entity detection, linking, relation prediction, and evidence combination, finding that basic LSTMs/GRUs with heuristics approach state-of-the-art accuracy on the SimpleQuestions dataset, while non-neural methods also perform reasonably well.
We examine the problem of question answering over knowledge graphs, focusing on simple questions that can be answered by the lookup of a single fact. Adopting a straightforward decomposition of the problem into entity detection, entity linking, relation prediction, and evidence combination, we explore simple yet strong baselines. On the popular SimpleQuestions dataset, we find that basic LSTMs and GRUs plus a few heuristics yield accuracies that approach the state of the art, and techniques that do not use neural networks also perform reasonably well. These results show that gains from sophisticated deep learning techniques proposed in the literature are quite modest and that some previous models exhibit unnecessary complexity.