US 11,960,989 B2
Read threshold estimation systems and methods using deep learning
Fan Zhang, Fremont, CA (US); Aman Bhatia, Los Gatos, CA (US); Xuanxuan Lu, San Jose, CA (US); Meysam Asadi, Fremont, CA (US); and Haobo Wang, San Jose, CA (US)
Assigned to SK hynix Inc., Gyeonggi-do (KR)
Filed by SK hynix Inc., Gyeonggi-do (KR)
Filed on Jul. 24, 2020, as Appl. No. 16/937,939.
Prior Publication US 2022/0027721 A1, Jan. 27, 2022
Int. Cl. G11C 29/00 (2006.01); G06F 3/06 (2006.01); G06N 3/08 (2023.01)
CPC G06N 3/08 (2013.01) [G06F 3/0604 (2013.01); G06F 3/0659 (2013.01); G06F 3/067 (2013.01)] 20 Claims
OG exemplary drawing
 
11. A method for operating a memory system, which includes a memory device including multiple pages coupled to select word lines in a memory region, and a controller coupled to the memory device, the method comprising:
performing multiple read operations on a select type of page for each word line using multiple read threshold sets;
obtaining fail bit count (FBC) information associated with the read operations, the FBC information including multiple FBC values corresponding to the multiple read threshold sets;
selecting a lowest FBC value among the multiple FBC values;
determining, using a neural network, a read threshold set corresponding to the lowest FBC value as an optimal read threshold set for each word line based on the FBC information, wherein the neural network utilizes suboptimal FBC information in addition to the optimal read threshold set to predict the optimal read threshold set; and
when optimal read threshold sets for the select word lines are different each other, predicting, using the neural network, a best read threshold set using the optimal read threshold sets.