US 11,914,600 B2
Multiple semantic hypotheses for search query intent understanding
Ming Wu, Kirkland, WA (US); Yong Ni, Bothell, WA (US); and Guo Mei, Lynnwood, WA (US)
Assigned to Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Jun. 30, 2021, as Appl. No. 17/363,208.
Prior Publication US 2023/0004568 A1, Jan. 5, 2023
Int. Cl. G06F 16/2457 (2019.01); G06N 20/00 (2019.01); G06N 5/04 (2023.01)
CPC G06F 16/24578 (2019.01) [G06N 5/041 (2013.01); G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
a processor; and
memory coupled to the processor, the memory comprising computer executable instructions that, when executed by the processor, performs a method comprising:
receiving a search query from a user;
identifying one or more tokens in the search query;
identifying one or more input signals associated with the one or more tokens, at least one input signal comprising an indication of an application entry point or a user access privilege;
generating a set of features using the one or more input signals;
generating one or more hypotheses based on the set of features, wherein the one or more hypotheses respectively represent a search query candidate for the search query and respectively comprise one or more of a domain, an intent, or a slot, wherein:
the domain indicates an area of knowledge or a data boundary;
the intent indicates a goal or an intention of the search query; and
the slot indicates actionable content within the search query;
ranking the one or more hypotheses based on at least the one or more input signals and a score assigned to a feature of the set of features of the one or more hypotheses, the score determined based on the least one input signal; and
processing the one or more ranked hypotheses.