US 11,893,991 B2
System and method for handling multi-turn conversations and context management for voice enabled ecommerce transactions
Snehasish Mukherjee, Santa Clara, CA (US); and Shankara Bhargava Subramanya, Santa Clara, CA (US)
Assigned to Walmart Apollo, LLC, Bentonville, AR (US)
Filed by Walmart Apollo, LLC, Bentonville, AR (US)
Filed on Jun. 24, 2022, as Appl. No. 17/848,636.
Application 17/848,636 is a continuation of application No. 16/177,274, filed on Oct. 31, 2018, granted, now 11,404,058.
Prior Publication US 2022/0319518 A1, Oct. 6, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 17/18 (2006.01); G10L 13/00 (2006.01); G10L 15/16 (2006.01); G10L 15/22 (2006.01); G10L 15/18 (2013.01)
CPC G10L 15/22 (2013.01) [G10L 15/1815 (2013.01); G10L 2015/223 (2013.01); G10L 2015/225 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A system, comprising:
a memory having instructions stored thereon; and
a processor configured to read the instructions to:
obtain, from a database, a context stack comprising a set of context entries, wherein each context entry in the set of context entries comprises a root intent element, an entity list element, and a dialogue stack, and wherein the dialogue stack of each context entry in the set of context entries includes a sequentially-stored structured data representation including least one user utterance and at least one system utterance, wherein each of the at least one user utterance and the at least one system utterance is stored as a data triplet including an intent variable, a set of entity variables and a timestamp;
receive machine-readable data indicative of a first user utterance from a user;
determine at least one existing semantic component, at least one missing semantic component, and an intent of the first user utterance;
identify a first context entry in the set of context entries based on the root intent elements of the set of context entries and the at least one existing semantic component of the first user utterance;
determine whether the at least one missing semantic component is identified within the dialogue stack of the first context entry;
in accordance with a determination that the at least one missing semantic component is identified within the dialogue stack of the first context entry and wherein the at least one missing semantic component includes at least one entity in the set of entities associated with the root intent element of the first context entry:
modify the sequentially-stored structured data representation of the first context entry to include a data triplet of the first user utterance,
generate an intent flow execution request including the at least one missing semantic component; and
in accordance with a determination that the at least one missing semantic component is unidentified within the dialogue stack of the first context entry:
create a new context entry comprising a new dialogue stack that includes a new sequentially-stored structured data representation,
modify the new sequentially-stored structured data representation of the new context entry to include a data triplet of the first user utterance,
generate, in response to the first user utterance, a first system utterance requesting the user to provide the at least one missing semantic component.