Neural link predictors are immensely useful for identifying missing edges in large scale Knowledge Graphs. However, it is still not clear how to use these models for answering more complex queries that arise in a number of domains, such as queries using logical conjunctions, disjunctions, and existential quantifiers, while accounting for missing edges. In this work, we propose a framework for efficiently answering complex queries on incomplete Knowledge Graphs. We translate each query into an end-to-end differentiable objective, where the truth value of each atom is computed by a pre-trained neural link predictor. We then analyse two solutions to the optimisation problem, including gradient-based and combinatorial search. In our experiments, the proposed approach produces more accurate results than state-of-the-art methods -- black-box neural models trained on millions of generated queries -- without the need of training on a large and diverse set of complex queries. Using orders of magnitude less training data, we obtain relative improvements ranging from 8% up to 40% in Hits@3 across different knowledge graphs containing factual information. Finally, we demonstrate that it is possible to explain the outcome of our model in terms of the intermediate solutions identified for each of the complex query atoms.

中文翻译：

---

使用神经链接预测器进行复杂查询回答

神经链接预测器对于识别大规模知识图中的缺失边非常有用。然而，仍然不清楚如何使用这些模型来回答在许多领域中出现的更复杂的查询，例如使用逻辑连词、析取和存在量词的查询，同时考虑缺失的边缘。在这项工作中，我们提出了一个有效回答不完整知识图上复杂查询的框架。我们将每个查询转换为端到端的可微分目标，其中每个原子的真值由预训练的神经链接预测器计算。然后我们分析了优化问题的两种解决方案，包括基于梯度和组合搜索。在我们的实验中，所提出的方法比最先进的方法（在数百万个生成的查询上训练的黑盒神经模型）产生更准确的结果，而无需对大量不同的复杂查询进行训练。使用较少数量级的训练数据，我们在包含事实信息的不同知识图中在 Hits@3 中获得了 8% 到 40% 的相对改进。最后，我们证明可以根据为每个复杂查询原子确定的中间解决方案来解释我们模型的结果。我们在包含事实信息的不同知识图中在 Hits@3 中获得了 8% 到 40% 的相对改进。最后，我们证明可以根据为每个复杂查询原子确定的中间解决方案来解释我们模型的结果。我们在包含事实信息的不同知识图中在 Hits@3 中获得了 8% 到 40% 的相对改进。最后，我们证明可以根据为每个复杂查询原子确定的中间解决方案来解释我们模型的结果。