US 12,079,571 B2
Identifying chat correction pairs for training model to automatically correct chat inputs
Saab Mansour, San Jose, CA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Jan. 27, 2023, as Appl. No. 18/160,960.
Application 18/160,960 is a continuation of application No. 17/030,204, filed on Sep. 23, 2020, granted, now 11,568,135.
Prior Publication US 2023/0177263 A1, Jun. 8, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 40/232 (2020.01); G06N 5/04 (2023.01); G06N 20/00 (2019.01); G06Q 30/0601 (2023.01); G06Q 50/00 (2024.01); H04L 51/212 (2022.01)
CPC G06F 40/232 (2020.01) [G06N 5/04 (2013.01); G06N 20/00 (2019.01); G06Q 30/0623 (2013.01); G06Q 30/0633 (2013.01); G06Q 50/01 (2013.01); H04L 51/212 (2022.05)] 20 Claims
OG exemplary drawing
 
1. A system, comprising:
one or more processors; and
one or more memories, wherein the one or more memories have stored thereon instructions, which when executed by the one or more processors, cause the one or more processors to:
identify one or more chat correction pairs that respectively comprise two chat inputs identified from among a plurality of previously obtained chat inputs that correspond to inputs from one or more users for one or more applications, wherein a given chat correction pair comprises two of the plurality of chat inputs from a given user, and wherein the given chat correction pair does not include one or more remaining chat inputs of the plurality of chat inputs, and wherein the given chat correction pair comprises:
a first chat input that has one or more errors; and
a second chat input that is a corrected version of the first chat input; and
subsequent to the identification of the one or more chat correction pairs, provide the one or more chat correction pairs to an endpoint.