Shaping Answers with Rules through Conversation

Dataset distribution under CC BY-SA 3.0

Data Visualisation

Data visualiser helps to understand the data by showing an instance in the form of a conversation. For an utterance, we also display its corresponding decision tree based on its tree_id. For a pair of snippet and a question, a decision tree shows all different paths to different answers, depending on the user context.

To view a specific instance, enter an 'utterance_id' from an instance in the training data and press View or press I'm Feeling Lucky to display a random instance.