Abstract
Being able to recognize words as slots and detect the intent of an utterance has been a keen
issue in natural language understanding. The
existing works either treat slot filling and intent detection separately in a pipeline manner, or adopt joint models which sequentially
label slots while summarizing the utterancelevel intent without explicitly preserving the
hierarchical relationship among words, slots,
and intents. To exploit the semantic hierarchy
for effective modeling, we propose a capsulebased neural network model which accomplishes slot filling and intent detection via a
dynamic routing-by-agreement schema. A rerouting schema is proposed to further synergize the slot filling performance using the inferred intent representation. Experiments on
two real-world datasets show the effectiveness
of our model when compared with other alternative model architectures, as well as existing
natural language understanding services.