US 12,346,228 B2
Methods and electronic device for repairing memory element in memory device
Helik Kanti Thacker, Bengaluru (IN); Adrita Barari, Bengaluru (IN); Akhilesh Sudhir Patankar, Bengaluru (IN); Deokgu Yoon, Suwon-si (KR); Damini, Bengaluru (IN); Keerthi Kiran Jagannathachar, Bengaluru (IN); Paulami Das, Bengaluru (IN); Sairam Jujjarapu, Bengaluru (IN); and Sudhanshu Gupta, Bengaluru (IN)
Assigned to SAMSUNG ELECTRONICS CO., LTD., Suwon-si (KR)
Filed by SAMSUNG ELECTRONICS CO., LTD., Suwon-si (KR)
Filed on Jul. 6, 2023, as Appl. No. 18/218,893.
Claims priority of application No. 202241039228 (IN), filed on Jul. 7, 2022; and application No. 202241039228 (IN), filed on Jun. 17, 2023.
Prior Publication US 2024/0012726 A1, Jan. 11, 2024
Int. Cl. G06F 11/00 (2006.01); G06F 11/22 (2006.01); G06N 3/04 (2023.01); G06N 3/092 (2023.01)
CPC G06F 11/2263 (2013.01) [G06N 3/04 (2013.01); G06N 3/092 (2023.01)] 15 Claims
OG exemplary drawing
 
1. A method comprising:
configuring, by an electronic device, a memory element as a graph with at least one vertex and at least one edge, at least one node associated with the memory element being encoded with information of at least one of a fault, a degree of the fault in a row of the memory element, a degree of the fault in a column of the memory element, a degree of spare rows available in the memory element, and a degree of spare columns available in the memory element;
determining, from the graph by the electronic device, a repair policy using a probability distribution over at least one of a faulty line and a non-faulty line as predicted by a graph neural network (GNN) based on a final node feature value from message passing stages of the GNN; and
determining, by the electronic device, a value of a state using a probability of the memory element being repaired from a particular state based on a global mean of all final node feature values predicted by the GNN.