Human languages are context-dependent by nature. There are many context-related system errors in current natural language understanding and generation models. This research considers contextual inform
The dataset used in the paper is a large database of query-reply pairs for retrieval-based conversation systems, containing 7 million query-reply pairs.
This dataset contains the model, dataset and training data used for training the CNN used in Spectrum Awareness winners of the Dyspan 2017 Challenge. If you use the this data for your own publications
