A Heterogeneous Graph with Factual, Temporal and Logical Knowledge for Question Answering Over Dynamic Contexts
This addresses the problem of interpretability and knowledge integration in neural QA systems, though it is incremental as it builds on existing graph-based methods.
The paper tackles question answering over dynamic textual contexts by proposing a heterogeneous graph that integrates factual, temporal, and logical knowledge, and shows that this approach improves a strong neural baseline on a benchmark dataset.
We study question answering over a dynamic textual environment. Although neural network models achieve impressive accuracy via learning from input-output examples, they rarely leverage various types of knowledge and are generally not interpretable. In this work, we propose a graph-based approach, where a heterogeneous graph is automatically built with factual knowledge of the context, temporal knowledge of the past states, and logical knowledge that combines human-curated knowledge bases and rule bases. We develop a graph neural network over the constructed graph, and train the model in an end-to-end manner. Experimental results on a benchmark dataset show that the injection of various types of knowledge improves a strong neural network baseline. An additional benefit of our approach is that the graph itself naturally serves as a rational behind the decision making.