US 12,282,517 B2
Memory system using heterogeneous data format and method of controlling the same
Won Woo Ro, Seoul (KR); Chanyoung Yoo, Seoul (KR); and Hongju Kal, Seoul (KR)
Assigned to UIF (University Industry Foundation), Yonsei University, Seoul (KR)
Filed by UIF (University Industry Foundation), Yonsei University, Seoul (KR)
Filed on Dec. 26, 2022, as Appl. No. 18/088,629.
Claims priority of application No. 10-2022-0018871 (KR), filed on Feb. 14, 2022.
Prior Publication US 2023/0259564 A1, Aug. 17, 2023
Int. Cl. G06F 16/95 (2019.01); G06F 16/9535 (2019.01); G06N 3/08 (2023.01)
CPC G06F 16/9535 (2019.01) [G06N 3/08 (2013.01)] 16 Claims
OG exemplary drawing
 
1. A memory system providing a personalized recommendation algorithm function to an internet service user based on a plurality of items, the memory system comprising:
a user preference analyzer for each item configured to calculate a user preference value corresponding to each item of an analysis target service;
a memory configured to store data related to the each item in a first data format or to store data related to the each item in a second data format with required bits less than the first data format, based on the user preference value of the each item;
an algorithm item extraction unit configured to extract a plurality of items from the analysis target service by using an artificial intelligence model;
an embedding table generating unit configured to generate an embedding table based on the extracted items; and
a personalized recommendation unit configured to perform the personalized recommendation algorithm function using the artificial intelligence model based on the embedding table, and to provide optimal personalized recommendation information to the internet service user,
wherein the algorithm item extraction unit is configured to extract a plurality of learning items from a machine learning service using the artificial intelligence model,
wherein the embedding table generation unit is configured to generate a learning embedding table based on the extracted learning items,
wherein the memory system further comprises a machine learning unit configured to train the artificial intelligence model by setting the learning embedding table as an input variable and setting the optimal personalized recommendation information as an output variable.