US 11,704,493 B2
Neural parser for snippets of dynamic virtual assistant conversation
Garfield W. Vaughn, South Windsor, CT (US); Gandhi Sivakumar, Bentleigh (AU); Vasanthi M. Gopal, Plainsboro, NJ (US); and Aaron K. Baughman, Cary, NC (US)
Assigned to KYNDRYL, INC., New York, NY (US)
Filed by KYNDRYL, INC., New York, NY (US)
Filed on Jan. 15, 2020, as Appl. No. 16/743,384.
Prior Publication US 2021/0216706 A1, Jul. 15, 2021
Int. Cl. G06F 40/30 (2020.01); G06F 40/205 (2020.01); G06N 3/02 (2006.01); G06F 18/24 (2023.01)
CPC G06F 40/30 (2020.01) [G06F 18/24 (2023.01); G06F 40/205 (2020.01); G06N 3/02 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method of pairing a user response and associated context with a neural network associated with a virtual assistant computer during a dynamic text conversation with an end user, the method comprising:
the virtual assistant computer receiving detected user generated text input;
the virtual assistant computer determining context of the detected user generated text input by:
employing a nonfunctional requirement compute component and a nonfunctional requirement scorer and ranker component; and
determining nonfunctional neural sentence component elements from the nonfunctional requirement compute component to compare the nonfunctional neural sentence component elements to a user pattern utterance ground truth database including a collection of micro level components of user behaviors and keyboard interactions to generate a nonfunctional score;
the virtual assistant computer comparing the context of the detected user generated text input by comparing a confidence score representing context of the user generated input to a classification associated with each of a plurality of existing nodes comprising a neural network;
for confidence scores below a threshold relative to the classification associated with each of the plurality of existing nodes of the neural network, the virtual assistant computer creating a new node within the neural network and assigning the context of the detected user generated text to the new node; and
for confidence scores equal to or above a threshold relative to the classification associated with each of the plurality of existing nodes of the neural network, the virtual assistant computer assigning the context of the detected user generated text to an existing node of the plurality of nodes of the neural network with a same or similar context.