Contextual spoken language understanding in a spoken dialogue system

A system for operating a spoken dialogue system, the system comprising:

• a processing unit;
• and
• a storage device for storing computer executable instructions that when performed by the processing unit, perform a method comprising:
• receiving, by the spoken dialogue system, a query;
• processing, by a spoken language apparatus in the spoken dialogue system, an input associated with the query to:
• analyze, using a neural network, the input and provide an output indicating the input is an out-of-domain query that is not associated with a domain that is absent in a predetermined set of domains supported by the spoken dialogue system;
• analyze, using the neural network, the input and provide an output indicating the input comprises a global command when the input is determined as the global command, the global command comprising an instruction to the spoken dialogue system to perform an action that is unrelated to a domain and relates to a change in a dialogue state;
• and
• analyze, using the neural network, the input and tag one or more entities in the input, each tag indicating:
• the entity that is tagged is to be used in determining an intent or a domain associated with the input;
• or
• the entity that is tagged is not to be used in determining the intent or the domain associated with the input.

A method of operating a spoken dialogue system that includes a global apparatus using a neural network, the method comprising:

• determining when a user query comprises one or more of:
• a global command, the global command comprising a command to the spoken dialogue system to perform an action that is unrelated to a domain and relates to a change in a dialogue state;
• an entity apparatus, the entity apparatus tagging one or more entities in the user query;
• an intent and domain apparatus, the intent and domain apparatus determining an intent and a domain associated with the user query;
• and
• a confirmation apparatus, the confirmation apparatus determining when the user query comprises a confirmation type;
• determining initial word and phrase embeddings using a skip-gram training process;
• adjusting the initial word and phrase embeddings using a plurality of lists to produce a first set of adjusted embeddings, each list comprising words and phrases that share similar semantic representations;
• training the spoken language understanding apparatus using the first set of adjusted embeddings;
• and
• adjusting the first set of adjusted embeddings to produce a second set of adjusted embeddings, wherein the first set of adjusted embeddings are adjusted to improve a correctness of an output produced by at least one of the global apparatus, the entity apparatus, the intent and domain apparatus, or the confirmation apparatus.