US 12,145,064 B2
Using data from a game metadata system to create actionable in-game decisions
Charles Denison, San Mateo, CA (US)
Assigned to Sony Interactive Entertainment Inc., Tokyo (JP)
Filed by Sony Interactive Entertainment Inc., Tokyo (JP)
Filed on Sep. 12, 2021, as Appl. No. 17/472,650.
Prior Publication US 2023/0082732 A1, Mar. 16, 2023
Int. Cl. A63F 13/533 (2014.01); A63F 13/67 (2014.01); A63F 13/79 (2014.01); A63F 13/822 (2014.01); G06N 20/00 (2019.01)
CPC A63F 13/533 (2014.09) [A63F 13/67 (2014.09); A63F 13/79 (2014.09); A63F 13/822 (2014.09); G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
at least one computer medium that is not a transitory signal and that comprises instructions executable by at least one processor to:
train at least a first machine learning (ML) model on plural computer game strategies, at least some of the strategies comprising at least one computer game character executing at least one activity using at least one mechanic in at least one computer game location, the at least some of the strategies being associated with ground truth outcomes;
subsequent to training, input to the ML model a current game information;
use the ML model to output an advisory to a player of a computer game generating the current game information regarding changing one or more of a computer game character, a mechanic, a location, an activity; and
present the advisory on at least one computer display, wherein presenting the advisory depends on a difficulty level such that a lower difficulty level results in presenting an advisory more often than an advisory is presented for a higher difficulty level.