US 12,265,544 B2
Domain knowledge guided selection of nodes for addition to data trees
Xiaoxuan Zhang, San Francisco, CA (US); Sijia Zhang, San Francisco, CA (US); and Yen-Yun Yu, Murray, UT (US)
Assigned to Ancestry.com Operations Inc., Lehi, UT (US)
Filed by Ancestry.com Operations Inc., Lehi, UT (US)
Filed on Jun. 3, 2022, as Appl. No. 17/831,439.
Application 17/831,439 is a continuation of application No. PCT/IB2022/054655, filed on May 19, 2022.
Claims priority of provisional application 63/191,192, filed on May 20, 2021.
Prior Publication US 2022/0382770 A1, Dec. 1, 2022
Int. Cl. G06F 15/16 (2006.01); G06F 16/2457 (2019.01); G06N 20/00 (2019.01)
CPC G06F 16/24578 (2019.01) [G06N 20/00 (2019.01)] 17 Claims
OG exemplary drawing
 
1. A computer-implemented method for continuously updating a set of nodes addable to a family tree, the computer-implemented method comprising:
generating for display, in a graphical user interface, an initial set of nodes that are addable to the family tree associated with a user;
receiving one or more selections from the user to interact with one or more genealogy nodes from the user interacting with the graphical user interface, a genealogy node corresponding to a genealogy data entry that is addable to the family tree; and
generating a continuously updated set of nodes for display in the graphical user interface, the continuously updated set of nodes being recommendations of genealogy data entries based on user interactions of genealogy data series as the family tree is displayed to the user, wherein the continuously updated set of nodes is updated based on past interactions of the user with one or more nodes, wherein generating the continuously updated set of nodes comprises:
tracking one or more last user interactions of the user performed at the graphical user interface;
updating, based on the one or more last user interactions performed at the graphical user interface, a recently interacted set of one or more interacted nodes with which the user has interacted within a number of past interactions;
selecting a pool of candidate nodes based on the recently interacted set, wherein at least one of the candidate nodes is within a domain boundary of one of the interacted nodes that is in the recently interacted set, and the domain boundary is defined at least based on a degree of kinship between a candidate node and an interacted node defined in a second family tree,
wherein selecting the pool of candidate nodes based on the recently interacted set is performed by a reinforcement machine learning model, selecting the pool of candidate nodes comprising:
collecting a plurality of candidate nodes that are within the domain boundary of at least one of the interacted nodes in the recently interacted set;
inputting a first set of features associated with the user to the reinforcement machine learning model, wherein the first set of features comprises user-level features and genealogical tree-level features;
inputting a second set of features representing the plurality of candidate nodes to the reinforcement machine learning model;
ranking the candidate nodes using the reinforcement machine learning model based on the first and second sets of features; and
selecting one or more candidate nodes of the plurality of candidate nodes as the pool of candidate nodes based on the ranking;
presenting, at the user interface and responsive to the one or more last user interactions performed at the graphical user interface, one or more candidate nodes of the pool of candidate nodes as a version of the continuously updated set of nodes; and
updating the pool of candidate nodes to be displayed in the graphical user interface as additional interactions that are performed by the user at the graphical user interface updates the recently interacted set.