US 12,282,500 B1
Providing completion recommendation variations for a partial natural language request
Dipock Das, Sunnyvale, CA (US); Dayanand Pochugari, Santa Clara, CA (US); and Aungon Nag Radon, Toronto (CA)
Assigned to Cisco Technology, Inc., San Jose, CA (US)
Filed by SPLUNK INC., San Francisco, CA (US)
Filed on Oct. 12, 2022, as Appl. No. 17/964,808.
Application 17/964,808 is a continuation of application No. 16/147,428, filed on Sep. 28, 2018, granted, now 11,475,053.
This patent is subject to a terminal disclaimer.
Int. Cl. G06F 16/332 (2025.01); G06F 3/0482 (2013.01); G06F 16/33 (2025.01); G06F 16/334 (2025.01); G06F 40/205 (2020.01); G06F 40/274 (2020.01); G06N 20/00 (2019.01)
CPC G06F 16/3322 (2019.01) [G06F 16/3344 (2019.01); G06F 40/205 (2020.01); G06F 40/274 (2020.01); G06N 20/00 (2019.01); G06F 3/0482 (2013.01)] 17 Claims
OG exemplary drawing
 
1. A computer-implemented method, comprising:
receiving a first incomplete natural language (NL) request and a first context associated with the first incomplete NL request;
generating a first set of completion recommendations based on the first incomplete NL request, the first context, and a first request completion model, wherein the first request completion model determines, for a given character sequence, a likelihood that the given character sequence follows one or more text strings included in the first incomplete NL request;
mapping a first request completion recommendation, included in the first set of request completion recommendations, to an intent;
generating, based on the intent, a second set of request completion recommendations, wherein each request completion recommendation included in the second set of request completion recommendations is generated by creating a variation in the wording of the first request completion recommendation;
receiving a selection of a first completion recommendation included in the first set of request completion recommendations or the second set of request completion recommendations as a selected completion recommendation;
generating a complete request based on the first incomplete NL request and the selected completion recommendation;
causing the complete request to be applied to a data storage system to generate a search result; and
presenting the search result based on a presentation recommendation, the presentation recommendation indicating a presentation format selected from a plurality of presentation formats in which the search result can be presented.