US 11,842,204 B2
Automated generation of early warning predictive insights about users
Shay Ben-Elazar, Herzliya (IL); Daniel Sitton, Tel Aviv (IL); Yossef Ben David, Tel Aviv (IL); Amnon Catav, Tel Aviv (IL); Meitar Ronen, Tel Aviv (IL); and Ori Bar-Ilan, Tel Aviv (IL)
Assigned to MICROSOFT TECHNOLOGY LICENSING, LLC, Redmond, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on May 3, 2021, as Appl. No. 17/306,847.
Claims priority of provisional application 63/166,562, filed on Mar. 26, 2021.
Prior Publication US 2022/0308895 A1, Sep. 29, 2022
Int. Cl. G06F 9/44 (2018.01); G06F 9/445 (2018.01); G06N 20/10 (2019.01); G06F 9/50 (2006.01); G06F 18/214 (2023.01)
CPC G06F 9/4451 (2013.01) [G06F 9/5072 (2013.01); G06F 18/214 (2023.01); G06N 20/10 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for automatically generating predictive insights from analysis of user activity data associated with an educational software platform, comprising:
accessing user activity data pertaining to user interactions by a plurality of users with the educational software platform, wherein the user activity data is raw signal data received from a plurality of applications or services associated with the educational software platform;
automatically applying a trained artificial intelligence (AI) model that is adapted to generate predictive insights from contextual relevance analysis of the user activity data, wherein the trained AI model executes processing operations that comprise:
generating a plurality of mobility determinations that identify changes in patterns of user behavior over a current temporal filter associated with the user activity data, wherein the plurality of mobility determinations are generated based on a collective relevance analysis that correlates:
data pertaining to current user interactions of a first user that are identified within the current temporal filter,
data pertaining to historical user interactions of the first user that are identified within a historical temporal filter associated with the user activity data,
data pertaining to current user interactions of one or more other users that are identified within the current temporal filter, and
data pertaining to historical user interactions of the one or more other users that are identified within the historical temporal filter, and
curating the plurality of mobility determinations to generate a curated listing of mobility determinations derived based on an application of business logic rules, configured for an education domain of the educational software platform, that are used to evaluate a relevance of the mobility determinations, and
generating one or more of the predictive insights based on an analysis of the curated listing of the plurality of mobility determinations; and
transmitting, to an application or service of the educational software platform, data for rendering the one or more predictive insights in a graphical user interface (GUI) notification displayable through the application or service of the educational software platform.