Power estimation using input vectors and deep recurrent neural networks
A method includes generating a plurality of input vectors based on input signals to an electric circuit, selecting a subset of the plurality of input vectors, and determining a plurality of datapoints based on the selected subset of the plurality of input vectors. Each datapoint of the plurality of datapoints indicates a power consumption of the electric circuit corresponding to an input vector of the selected subset of the input vectors. The method also includes generating, by a processor, a plurality of vector sequences based on the selected subset of the plurality of input vectors. Each vector sequence of the plurality of vector sequences includes a portion of the selected subset of the plurality of input vectors arranged chronologically. The method further includes training a machine learning model based on a first subset of the plurality of vector sequences and a corresponding first subset of the plurality of datapoints.
1. A method comprising:
generating a plurality of input vectors, wherein each input vector of the plurality of input vectors indicates values of input signals provided to input ports of an electric circuit during a cycle period of the electric circuit;
generating a plurality of datapoints by simulating, for each input vector of the plurality of input vectors, the electric circuit to determine a power consumption of the electric circuit when the values of the input signals indicated by that input vector are input to the electric circuit;
generating, by a processor, a plurality of vector sequences by generating, for each input vector of the plurality of input vectors, a vector sequence comprising that input vector and a set of input vectors arranged chronologically according to times when that input vector and the set of input vectors were generated, wherein the set of input vectors were generated prior to that input vector, and wherein the vector sequence corresponds to the power consumption of the electric circuit for that input vector; and
training a machine learning model based on a first subset of the plurality of vector sequences and a first subset of the plurality of datapoints corresponding to the first subset of the plurality of vector sequences.
2. The method of claim 1 , further comprising reducing a dimension of the plurality of vector sequences using principal component analysis or multiple correspondence analysis.
3. The method of claim 1 , further comprising dividing the plurality of vector sequences and the plurality of datapoints into the first subset of the plurality of vector sequences, the corresponding first subset of the plurality of datapoints, a second subset of the plurality of vector sequences, and a corresponding second subset of the plurality of datapoints.
4. The method of claim 3 , further comprising validating the machine learning model using the second subset of the plurality of vector sequences and the corresponding second subset of the plurality of datapoints.
5. The method of claim 1 , further comprising:
monitoring a validation error while training the machine learning model; and
stopping the training of the machine learning model when the validation error is not reduced after some epochs of the training.
6. The method of claim 1 , wherein the machine learning model is a Long Short-Term Memory model.
7. The method of claim 1 , wherein the machine learning model is a Gated Recurrent Unit model.
8. The method of claim 1 , wherein a vector sequence of the plurality of vector sequences comprises an internal vector generated based on an internal signal of the electric circuit.
9. A computer system comprising:
a memory; and
a hardware processor communicatively coupled to the memory, the hardware processor configured to:
generate a plurality of input vectors, wherein each input vector of the plurality of input vectors indicates values of input signals provided to input ports of an electric circuit during a cycle period of the electric circuit;
generate a plurality of datapoints by simulating, for each input vector of the plurality of input vectors, the electric circuit to determine a power consumption of the electric circuit when the values of the input signals indicated by that input vector are input to the electric circuit;
generate a plurality of vector sequences by generating, for each input vector of the plurality of input vectors, a vector sequence comprising that input vector and a set of input vectors arranged chronologically according to times when that input vector and the set of input vectors were generated, wherein the set of input vectors were generated prior to that input vector, and wherein the vector sequence corresponds to the power consumption of the electric circuit for that input vector; and
train a machine learning model based on a first subset of the plurality of vector sequences and a corresponding first subset of the plurality of datapoints corresponding to the first subset of the plurality of vector sequences.
10. The computer system of claim 9 , wherein the hardware processor is further configured to reduce a dimension of the plurality of vector sequences using principal principle component analysis or multiple correspondence analysis.
11. The computer system of claim 9 , wherein the hardware processor is further configured to divide the plurality of vector sequences and the plurality of datapoints into the first subset of the plurality of vector sequences, the corresponding first subset of the plurality of datapoints, a second subset of the plurality of vector sequences, and a corresponding second subset of the plurality of datapoints.
12. The computer system of claim 11 , wherein the hardware processor is further configured to validate the machine learning model using the second subset of the plurality of vector sequences and the second subset of the plurality of datapoints.
13. The computer system of claim 9 , wherein the hardware processor is further configured to:
monitor a validation error while training the machine learning model; and
stop training the machine learning model when the validation error is not reduced after some epochs of the training.
14. The computer system of claim 9 , wherein the machine learning model is a Long Short-Term Memory model.
15. The computer system of claim 9 , wherein the machine learning model is a Gated Recurrent Unit model.
16. The computer system of claim 9 , wherein a vector sequence of the plurality of vector sequences comprises an internal vector generated based on an internal signal of the electric circuit.
17. A method comprising:
generating a plurality of input vectors, wherein each input vector of the plurality of input vectors indicates values of input signals provided to input ports of an electric circuit during a cycle period of the electric circuit;
generating, by a processor, a plurality of vector sequences by generating, for each input vector of the plurality of input vectors, a vector sequence comprising that input vector and a set of input vectors arranged chronologically according to times when that input vector and the set of input vectors were generated, wherein the set of input vectors were generated prior to that input vector, and wherein the vector sequence corresponds to power consumption of the electric circuit for that input vector; and
applying a machine learning model to the plurality of vector sequences to predict a power consumption of the electric circuit.
18. The method of claim 17 , wherein the machine learning model is a Long Short-Term Memory model.