公开(公告)号：US20220405588A1

公开(公告)日：2022-12-22

申请号：US17824556

申请日：2022-05-25

Applicant: Ishaan KUMAR ,  Yaochen HU ,  Yingxue ZHANG

Inventor： Ishaan KUMAR ,  Yaochen HU ,  Yingxue ZHANG

IPC: G06N3/08 ,  G06N3/04 ,  G06N3/063

Abstract: Systems, methods, and computer-readable media provide a graph processing system that incorporates a graph neural network (GNN) based recommender system (RS), as well as a method for training a GNN based RS to address feature leakage that leads to overfitting of the trained GNN based RS. A message correction algorithm is used to modify a user node embedding and a positive item node embedding generated by the graph neural network when generating mini batches of training triples used to train the GNN based RS. The GNN message passing operations are performed on one graph only, in contrast to existing approaches which typically run GNN message passing operations on multiple adjusted input graphs constructed for multiple training triples.

公开(公告)号：US20240169132A1

公开(公告)日：2024-05-23

申请号：US17991806

申请日：2022-11-21

Applicant: Surya PENMETSA ,  Yingying FU ,  Yingxue ZHANG ,  Yaochen HU

Inventor： Surya PENMETSA ,  Yingying FU ,  Yingxue ZHANG ,  Yaochen HU

IPC: G06F30/323 ,  G06N3/0464

CPC classification number: G06F30/323 ,  G06N3/0464

Abstract: Systems and methods for selecting one or more characteristics for a netlist. A graph data structure is generated, the components of the netlist and connections therebetween being represented by nodes and edges, respectively, in the graph data structure, wherein the edges between the nodes indicate types of the connections between the components. The graph data structure is processed to select one or more characteristics for the components.