Adapting the DisCoCat-Model for Question Answering in the Chinese Language

We introduce quantum natural language processing for the Chinese language. Our approach focuses on Question-answering on a set of sentences, whether a particular sentence is truthful with respect to the whole corpus of sentences. We employ the Categorical Distributional Compositional (DisCoCat) model to translate sentences to valid quantum circuits. We achieve a fitting score of 97% on the test set. Our sentence set is also significantly larger than previous experiments. The results show general applicability of the framework to other languages. We also show that it can be used to introspect natural language models and provide new approaches to model explainability.