Storage device inferring read levels based on artificial neural network model and learning method of artificial neural network model

A storage device comprising:

• a non-volatile memory including a plurality of blocks;
• a buffer memory configured to store an artificial neural network model;
• and
• a controller configured to:
• control the non-volatile memory and the buffer memory,
• read memory cells connected to reference word lines of the plurality of blocks by using first and second read levels for acquiring a plurality of first and second on-cell counts of the plurality of blocks,
• store the plurality of first and second on-cell counts of the plurality of blocks in the buffer memory,
• input first and second on-cell counts corresponding to a target block among the plurality of first and second on-cell counts and a number of a target word line of the target block to the artificial neural network model, and
• infer a plurality of read levels for reading data of memory cells connected to the target word line using the artificial neural network model.

A method of learning read levels to be applied to word lines of each of a plurality of blocks in a non-volatile memory of a storage device by using an artificial neural network model, the method comprising:

• reading memory cells connected to a reference word line of each of the plurality of blocks by using a read level to acquire an on-cell count of each of the plurality of blocks;
• calculating a plurality of read levels for reading data of memory cells connected to a plurality of word lines of each of the plurality of blocks;
• providing the artificial neural network model with training data, which comprise input data comprising the on-cell count of each of the plurality of blocks and numbers of the plurality of word lines and output data comprising the plurality of read levels;
• and
• calibrating weights and biases of the artificial neural network model such that the artificial neural network model generates the output data from the input data.