TASK SWITCHING IN DIALOGUE PROCESSING

摘要

Disclosed methods and systems are directed to task switching in dialogue processing. The methods and systems may include activating a primary task, receiving, one or more ambiguous natural language commands, and identifying a first candidate task for each of the one or more ambiguous natural language commands. The methods and system may also include identifying, for each of the one or more ambiguous natural language commands and based on one or more rules, a second candidate task of the plurality of tasks corresponding to the ambiguous natural language command, determining whether to modify at least one of the one or more rules-based task switching rules based on whether a quality metric satisfies a threshold quantity, and when the second quality metric satisfies the threshold quantity, changing the task switching rule for the corresponding candidate task from a rules-based model to the optimized statistical based task switching model.

主权项

1. A method comprising:
activating, by a computing device and with an application comprising a plurality of tasks, a primary task in response to receiving a first natural language command of one or more words, the computing device being configurable with one or more rules-based task switching rules for determining one or more secondary tasks, the secondary tasks comprising tasks that cannot be activated while one or more primary tasks are activated and tasks that can be activated while one or more primary tasks are activated; receiving, while the primary task is activated, one or more ambiguous natural language commands; determining at least one candidate task of the plurality of tasks for each of the one or more ambiguous natural language commands; assigning a score to each of the candidate tasks using a statistical-based task switching model; identifying, based on the assigned scores, a first candidate task of the one or more candidate tasks for each of the one or more ambiguous natural language commands; identifying, for each of the one or more ambiguous natural language commands and based on the one or more rules of the rules-based task switching rules, a second candidate task of the plurality of tasks corresponding to the ambiguous natural language command; identifying an optimal task of the plurality of tasks for each of the one or more ambiguous natural language commands; optimizing the statistical task switching model based on a first quality metric, the first quality metric being based on the identified optimal tasks; calculating a second quality metric based on whether at least a portion of the identified first candidate tasks correspond to the secondary tasks; determining whether to modify at least one of the one or more rules-based task switching rules based on whether the second quality metric satisfies a threshold quantity; and when the second quality metric satisfies the threshold quantity, changing the task switching rule for the corresponding candidate task from a rules-based model to the optimized statistical based task switching model.