Joint intent detection and slot filling with wheel-graph attention networks

Intent detection and slot filling are recognized as two very important tasks in a spoken language understanding (SLU) system. In order to model these at the same time, many joint models based on deep neural networks have been proposed recently archived excellent results. addition, graph network has made good achievements field of vision. Therefore, we combine advantages propose new with wheel-graph attention (Wheel-GAT), which is able interrelated connections directly for single intent filling. To construct structure utterances, create nodes, directed edges. nodes can provide utterance-level semantic information filling, while also local keyword detection. The promote each other carry out end-to-end training time. Experiments show that our approach superior multiple baselines ATIS SNIPS datasets. Besides, demonstrate using bi-directional encoder representation from transformer (BERT) further boosts performance SLU task.