Read threshold voltage estimation systems and methods for parametric PV-level modeling
Embodiments provide a scheme for estimating an optimal read threshold voltage using a deep neural network (DNN) with reduced number of processing. A controller receives first and second program voltage (PV) levels associated with read operations on cells. The controller estimates first and second probability distribution parameter sets representing skew normal distributions of the first and second PV levels, respectively. The controller estimates an optimal read threshold voltage based on the first and second probability distribution parameter sets. The optimal read threshold voltage is a read threshold voltage such that first probability density function (PDF) value of the skew normal distribution of the first PV level is the same as the second PDF value of the skew normal distribution of the second PV level.
1. A memory system comprising:
a memory device including a plurality of cells; and
a controller configured to:
receive first and second program voltage (PV) levels associated with read operations on the cells;
determine skew normal distributions for the first and second PV levels, respectively;
estimate, by a first neural network, first and second probability distribution parameter sets based on the skew normal distributions of the first and second PV levels, respectively;
determine, by a second neural network, first probability density function (PDF) values and second PDF values for respective ones of the first and second probability distribution parameter sets;
estimate an optimal read threshold voltage based on the first and second PDF values, the optimal read threshold voltage being a read threshold voltage that is determined based on a probability density function value of the first PDF values that is approximately equal to a probability density function value of the second PDF values, and
perform a read operation for the memory device based on the optimal read threshold voltage.
2. The memory system of claim 1 , wherein the controller is further configured to perform a next read operation on the cells using the optimal read threshold voltage.
3. The memory system of claim 1 , wherein the first and second probability distribution parameter sets include location, scale and shape of a curve associated with the skew normal distributions of the first and second PV levels, respectively.
4. The memory system of claim 1 , wherein the optimal read threshold voltage is estimated as the cross-point voltage of the first and second PDF values.
5. The memory system of claim 1 , wherein the neural network performs fixed-point multiply-accumulate (MAC) operations.
6. The memory system of claim 1 , wherein each of the first and second probability distribution parameter sets is estimated by the first neural network, which is trained to output each of multiple probability distribution parameter sets corresponding to each of the first and second PV levels.
7. The memory system of claim 6 , wherein the second neural network performs fixed-point multiply-accumulate (MAC) operations.
8. A method for operating a memory system including a memory device including a plurality of cells and a controller, the method comprising:
receiving first and second program voltage (PV) levels associated with read operations on the cells;
determining skew normal distributions for the first and second PV levels, respectively;
estimating, by a first neural network, first and second probability distribution parameter sets based on skew normal distributions of the first and second PV levels, respectively;
determining, by a second neural network, first probability density function (PDF) values and second PDF values for respective ones of the first and second probability distribution parameter sets;
estimating an optimal read threshold voltage based on the first and second PDF values, the optimal read threshold voltage being a read threshold voltage that is determined based on a probability density function value of the first PDF values that is approximately equal to a probability density function value of the second PDF values, and
performing a read operation for the memory device based on the optimal read threshold voltage.
9. The method of claim 8 , wherein further comprising:
performing a next read operation on the cells using the optimal read threshold voltage.
10. The method of claim 8 , wherein the first and second probability distribution parameter sets include location, scale and shape of a curve associated with the skew normal distributions of the first and second PV levels, respectively.
11. The method of claim 8 , wherein the optimal read threshold voltage is estimated as the cross-point voltage of the first and second PDF values.
12. The method of claim 8 , wherein the second neural network performs fixed-point multiply-accumulate (MAC) operations.
13. The method of claim 8 , wherein each of the first and second probability distribution parameter sets is estimated by the first neural network, which is trained to output each of multiple probability distribution parameter sets corresponding to each of the first and second PV levels.
14. The method of claim 13 , wherein the second neural network performs fixed-point multiply-accumulate (MAC) operations.