US 11,868,720 B2
System for multi-perspective discourse within a dialog
Vivek Varma Datla, Ashland, MA (US); Sheikh Sadid Al Hasan, Cambridge, MA (US); Aaditya Prakash, Waltham, MA (US); Oladimeji Feyisetan Farri, Yorktown Heights, NY (US); Tilak Raj Arora, Gurgaon (IN); Junyi Liu, Windham, NH (US); and Ashequl Qadir, Melrose, MA (US)
Assigned to KONINKLIJKE PHILIPS N.V., Eindhoven (NL)
Appl. No. 17/420,750
Filed by KONINKLIJKE PHILIPS N.V., Eindhoven (NL)
PCT Filed Jan. 16, 2020, PCT No. PCT/EP2020/050954
§ 371(c)(1), (2) Date Jul. 6, 2021,
PCT Pub. No. WO2020/148355, PCT Pub. Date Jul. 23, 2020.
Claims priority of provisional application 62/793,611, filed on Jan. 17, 2019.
Prior Publication US 2022/0108068 A1, Apr. 7, 2022
Int. Cl. G06F 40/279 (2020.01); G06N 20/00 (2019.01)
CPC G06F 40/279 (2020.01) [G06N 20/00 (2019.01)] 15 Claims
OG exemplary drawing
 
15. A system comprising one or more processors and memory operably coupled with the one or more processors, wherein the memory stores instructions that, in response to execution of the instructions by one or more processors, cause the one or more processors to perform the following operations:
receiving user interface input captured by a computing system during a dialog between a user and a user-facing agent that executes at least in part on the computing system;
processing the user interface input using a set of sub-agents by processing the user interface input using a trained machine learning model for each sub-agent to generate a set of candidate responses, wherein each sub-agent and each corresponding trained machine learning model is tailored to a different dialog feature in a set of dialog features;
processing each candidate response in the set of candidate responses using a set of feature monitoring processes to generate a set of dialog feature values corresponding to the set of dialog features for each candidate response;
selecting a response to the user interface input from the set of candidate responses by comparing the set of dialog feature values for each candidate response to a set of global priority objectives for the dialog, wherein the set of global priority objectives include predetermined values for each feature in the set of dialog features;
updating one or more aspects of each sub-agent with the selected response; and
causing the computing system to render output from the user-facing agent based on the selected response.