Answering Binary Causal Questions Using Role-Oriented Concept Embedding

Answering binary causal questions is a challenging task, and it requires rich background knowledge to answer such questions. Extracting useful causal features from the background knowledge base and applying them effectively in a model is a crucial step to answering binary causal questions. The state-of-the-art approaches apply deep learning techniques to answer binary causal questions. In these approaches, candidate concepts are often embedded into vectors to model causal relationships among them. However, a concept may play the role of a cause in one question, but it could be an effect in another question. This aspect has not been extensively explored in existing approaches. Role-oriented causal concept embeddings are proposed in this article to model causality between concepts. We also propose leveraging semantic concept similarity to extract causal information from concepts. Finally, we develop a deep learning framework to answer binary causal questions. Our approach yields accuracy that is comparable to or better than the benchmark approaches.

I. Introduction

Binary questions are frequently asked for the confirmation of given information. These questions can be answered by yes/no answers. Similarly, binary causal questions (BCQs) are asked to confirm whether or not a causal relationship exists between two candidate concepts. Fig. 1 depicts an example BCQ. In the example question, “Could COVID-19 cause lung failure?,” COVID-19 and lung failure are two candidate concepts, and the question asks whether or not there is a causal relationship between them.