| CPC G10L 15/07 (2013.01) [G06F 18/2413 (2023.01); G06N 20/00 (2019.01); G06Q 30/0631 (2013.01); G10L 15/083 (2013.01); G10L 2015/088 (2013.01)] | 17 Claims |
|
1. A server comprising:
one or more processors; and
one or more non-transitory computer readable storage media to store instructions executable by the one or more processors to perform operations comprising:
receiving, by an automated voice ordering system associated with a restaurant that is executing on the server, speech comprising order data from a device associated with a customer, the automated voice ordering system based on a machine learning model comprising a neural network trained using a plurality of customer-employee conversations;
performing natural language processing of the speech to identify one or more items in a menu;
adding the one or more items to an order displayed on a point-of-sale terminal of the restaurant, the order associated with the customer, wherein after adding the one or more items to the order displayed on the point-of-sale terminal of the restaurant:
receiving a correction, from an employee via the point-of-sale terminal, to at least one item of the one or more items;
saving the correction in training data; and
training the neural network using the training data;
predicting, using the neural network, one or more suggested items based at least in part on a similarity graph of a plurality of menu items in the menu, location, and colloquial terms;
providing, by the neural network, the one or more suggested items to the device as part of a conversation between the neural network and the customer;
receiving, from the device, additional speech identifying at least one selection from the one or more suggested items;
adding the at least one selection to the order;
determining, by the neural network, that the order is complete;
initiating fulfillment of the order comprising providing an instruction to an employee of the restaurant to gather one or more physical items included in the order;
wherein the similarity graph is based on one or more tags associated with each of a plurality of menu items in the menu, the one or more tags comprising:
a dietary type tag associated with each of the plurality of menu items;
a nutritional type tag associated with each of the plurality of menu items;
an ingredient group tag associated with each of the plurality of menu items;
a dish type tag associated with each of the plurality of menu items;
a cooking method tag associated with each of the plurality of menu items;
or any combination thereof;
storing data associated with the conversation between the neural network and the customer to create stored data; and
after a predetermined period of time has elapsed, retraining the neural network using the stored data.
|