Systems and methods for modeless read threshold voltage estimation
Embodiments provide a scheme for estimating an optimal read threshold voltage using a deep neural network (DNN) with a reduced number of processing. A controller includes a combined neural network, which receives first and second cumulative distribution function (CDF) values, each CDF value corresponding to a program voltage (PV) level associated with a read operation on the cells. The combined neural network generates first and second connection vectors based on the first and second CDF values and first weight values, and estimates an optimal read threshold voltage based on the first and second connection vectors and second weight values.
1. A memory system comprising:
a memory device including a plurality of cells; and
a controller including a combined neural network configured to:
receive first and second cumulative distribution function (CDF) values, each CDF value corresponding to a program voltage (PV) level associated with a read operation on the cells;
generate first and second connection vectors based on the first and second CDF values and first weight values, the first connection vector including first values representing a threshold voltage distribution of a first PV level and the second connection vector including second values representing a threshold voltage distribution of a second PV level; and
estimate an optimal read threshold voltage based on the first and second connection vectors and second weight values,
wherein the combined neural network includes:
a first neural network configured to receive the first CDF value and generate the first connection vector based on the first CDF value and the first weight values;
a second neural network configured to receive the second CDF value and generate the second connection vector based on the second CDF value and the first weight values; and
a third neural network configured to receive the first and second connection vectors and generate the optimal read threshold voltage based on the first and second connection vectors and the second weight values.
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 combined neural network is trained on a set dataset including CDF values, sampling voltages corresponding to the CDF values and the optimal read threshold voltage.
4. The memory system of claim 3 , wherein the combined neural network measures the error between the output of the combined neural network and the optimal read threshold voltage to determine a set loss function.
5. The memory system of claim 4 , wherein the first and second weight values are obtained through a stochastic gradient descent such that the loss function is minimized.
6. The memory system of claim 4 , wherein the error includes a mean square error between the output of the combined neural network and the optimal read threshold voltage.
7. The memory system of claim 4 , wherein each weight values include multiple matrices and biases vectors.
8. The memory system of claim 7 , wherein the first neural network generates the first connection vector based on the first CDF value, the first weight values and a set activation function;
the second neural network generates the second connection vector based on the second CDF value, the first weight values and the activation function; and
the third neural network generates the optimal read threshold voltage based on the first and second connection vectors, the second weight values and the activation function.
9. A method for operating a memory system including a memory device including a plurality of cells and a controller including a combined neural network, the method comprising:
receiving first and second cumulative distribution function (CDF) values, each CDF value corresponding to a program voltage (PV) level associated with a read operation on the cells;
generating first and second connection vectors based on the first and second CDF values and first weight values, the first connection vector including first values representing a threshold voltage distribution of a first PV level and the second connection vector including second values representing a threshold voltage distribution of a second PV level; and
estimating an optimal read threshold voltage based on the first and second connection vectors and second weight values,
wherein the combined neural network includes a first neural network, a second neural network and a third neural network, and
wherein the estimating of the optimal read threshold includes:
receiving, by the first neural network, the first CDF value and generating the first connection vector based on the first CDF value and the first weight values;
receiving, by the second neural network, the second CDF value and generating the second connection vector based on the second CDF value and the first weight values; and
receiving, by the third neural network, the first and second connection vectors and generating the optimal read threshold voltage based on the first and second connection vectors and the second weight values.
10. The method of claim 9 , wherein further comprising:
performing a next read operation on the cells using the optimal read threshold voltage.
11. The method of claim 9 , further comprising:
training the combined neural network on a set dataset including CDF values, sampling voltages corresponding to the CDF values and the optimal read threshold voltage.
12. The method of claim 11 , further comprising:
measuring, by the combined neural network, the error between the output of the combined neural network and the optimal read threshold voltage to determine a set loss function.
13. The method of claim 12 , wherein the first and second weight values are obtained through a stochastic gradient descent such that the loss function is minimized.
14. The method of claim 12 , wherein the error includes a mean square error between the output of the combined neural network and the optimal read threshold voltage.
15. The method of claim 12 , wherein each weight values include multiple matrices and biases vectors.
16. The method of claim 15 , wherein the first neural network generates the first connection vector based on the first CDF value, the first weight values and a set activation function;
the second neural network generates the second connection vector based on the second CDF value, the first weight values and the activation function; and
the third neural network generates the optimal read threshold voltage based on the first and second connection vectors, the second weight values and the activation function.
17. A memory system comprising:
a memory device including a plurality of cells; and
a controller including a combined neural network configured to:
receive first and second cumulative distribution function (CDF) values, each CDF value corresponding to a program voltage (PV) level associated with a read operation on the cells;
generate first and second connection vectors based on the first and second CDF values and first weight values; and
estimate an optimal read threshold voltage based on the first and second connection vectors and second weight values,
wherein the combined neural network includes:
a first neural network configured to receive the first CDF value and generate the first connection vector based on the first CDF value and the first weight values;
a second neural network configured to receive the second CDF value and generate the second connection vector based on the second CDF value and the first weight values; and
a third neural network configured to receive the first and second connection vectors and generate the optimal read threshold voltage based on the first and second connection vectors and the second weight values.
18. A method for operating a memory system including a memory device including a plurality of cells and a controller including a combined neural network, the method comprising:
receiving first and second cumulative distribution function (CDF) values, each CDF value corresponding to a program voltage (PV) level associated with a read operation on the cells;
generating first and second connection vectors based on the first and second CDF values and first weight values; and
estimating an optimal read threshold voltage based on the first and second connection vectors and second weight values,
wherein the combined neural network includes a first neural network, a second neural network and a third neural network, and
wherein the estimating of the optimal read threshold includes:
receiving, by the first neural network, the first CDF value and generating the first connection vector based on the first CDF value and the first weight values;
receiving, by the second neural network, the second CDF value and generating the second connection vector based on the second CDF value and the first weight values; and
receiving, by the third neural network, the first and second connection vectors and generating the optimal read threshold voltage based on the first and second connection vectors and the second weight values.