Method for model adaptation, electronic device and computer program product

1. A method for model adaptation, comprising:

processing first input data by using a first machine learning model having first parameter set values, to obtain first feature information of the first input data, the first machine learning model having a capability of self-ordering and the first parameter set values being updated after the processing of the first input data;

generating a first classification result for the first input data based on the first feature information by using a second machine learning model having second parameter set values;

processing second input data by using the first machine learning model having the updated first parameter set values, to obtain second feature information of the second input data;

generating a second classification result for the second input data based on the second feature information by using the second machine learning model having the second parameter set values; and

generating adaptive change output data based on changes in the first input data and the second input data;

the method further comprising:

storing at least one of the first feature information and the second feature information; and

in accordance with a presence of at least one of a first ground-truth classification result for the first input data and a second ground-truth classification result for the second input data, and in accordance with a determination that an update of the second machine learning model is triggered, re-training a duplicated model of the second machine learning model by using at least one of a pair of the first feature information and the first ground-truth classification result and a pair of the second feature information and the second ground-truth classification result, so as to update the second parameter set values;

wherein the first machine learning model comprises an unsupervised spiking neural network; and

wherein the second machine learning model comprises a machine learning model other than an unsupervised spiking neural network and a supervised spiking neural network.

2. The method of claim 1 ,

wherein processing the first input data comprises: converting the first input data into first spiking stream data available for processing by the unsupervised spiking neural network, and

wherein processing the second input data comprises: converting the first input data into second spiking stream data available for processing by the unsupervised spiking neural network.

3. The method of claim 2 , wherein the first machine learning model is run on a neuromorphic chip.

4. The method of claim 1 , wherein the first parameter set values are obtained by training the first machine learning model based on training input data and ground-truth label data for the training input data.

5. The method of claim 1 , wherein the second parameter set values of the second machine learning model are obtained through a supervised learning process.

6. The method of claim 1 , further comprising:

in accordance with an absence of the first ground-truth classification result for the first input data or the second ground-truth classification result for the second input data for a predetermined time period, discarding the first feature information or the second feature information.

7. The method of claim 6 , further comprising:

configuring the second machine learning model with the updated second parameter set values.

8. An electronic device, comprising:

at least one processor; and

at least one memory storing computer program instructions, the at least one memory and the computer program instructions being configured, with the at least one processor, to cause the electronic device to perform acts comprising:

processing first input data by using a first machine learning model having first parameter set values, to obtain first feature information of the first input data, the first machine learning model having a capability of self-ordering and the first parameter set values being updated after the processing of the first input data;

generating a first classification result for the first input data based on the first feature information by using a second machine learning model having second parameter set values;

processing second input data by using the first machine learning model having the updated first parameter set values, to obtain second feature information of the second input data;

generating a second classification result for the second input data based on the second feature information by using the second machine learning model having the second parameter set values; and

generating adaptive change output data based on changes in the first input data and the second input data;

wherein the acts further comprise:

storing at least one of the first feature information and the second feature information; and

in accordance with a presence of at least one of a first ground-truth classification result for the first input data and a second ground-truth classification result for the second input data, and in accordance with a determination that an update of the second machine learning model is triggered, re-training a duplicated model of the second machine learning model by using at least one of a pair of the first feature information and the first ground-truth classification result and a pair of the second feature information and the second ground-truth classification result, so as to update the second parameter set values;

wherein the first machine learning model comprises an unsupervised spiking neural network; and

wherein the second machine learning model comprises a machine learning model other than an unsupervised spiking neural network and a supervised spiking neural network.

9. The electronic device of claim 8 , wherein processing the first input data comprises: converting the first input data into first spiking stream data available for processing by the unsupervised spiking neural network, and

wherein processing the second input data comprises: converting the first input data into second spiking stream data available for processing by the unsupervised spiking neural network.

10. The electronic device of claim 9 , wherein the first machine learning model is run on a neuromorphic chip.

11. The electronic device of claim 8 , wherein the first parameter set values are obtained by training the first machine learning model based on training input data and ground-truth label data for the training input data.

12. The electronic device of claim 8 , wherein the second parameter set values of the second machine learning model are obtained through a supervised learning process.

13. The electronic device of claim 8 , wherein the acts further comprise:

in accordance with an absence of the first ground-truth classification result for the first input data or the second ground-truth classification result for the second input data for a predetermined time period, discarding the first feature information or the second feature information.

14. The electronic device of claim 13 , wherein the acts further comprise:

configuring the second machine learning model with the updated second parameter set values.

15. A computer program product which is tangibly stored on a non-transitory computer readable medium and comprises machine-executable instructions which, when executed by a device, causing the device to:

process first input data by using a first machine learning model having first parameter set values, to obtain first feature information of the first input data, the first machine learning model having a capability of self-ordering and the first parameter set values being updated after the processing of the first input data;

generate a first classification result for the first input data based on the first feature information by using a second machine learning model having second parameter set values;

process second input data by using the first machine learning model having the updated first parameter set values, to obtain second feature information of the second input data;

generate a second classification result for the second input data based on the second feature information by using the second machine learning model having the second parameter set values; and

generate adaptive change output data based on changes in the first input data and the second input data;

wherein the machine-executable instructions, when executed by the device, further cause the device to:

store at least one of the first feature information and the second feature information; and

in accordance with a presence of at least one of a first ground-truth classification result for the first input data and a second ground-truth classification result for the second input data, and in accordance with a determination that an update of the second machine learning model is triggered, re-train a duplicated model of the second machine learning model by using at least one of a pair of the first feature information and the first ground-truth classification result;

wherein the first machine learning model comprises an unsupervised spiking neural network; and

wherein the second machine learning model comprises a machine learning model other than an unsupervised spiking neural network and a supervised spiking neural network.

16. The computer program product of claim 15 , wherein processing the first input data comprises: converting the first input data into first spiking stream data available for processing by the unsupervised spiking neural network, and

wherein processing the second input data comprises: converting the first input data into second spiking stream data available for processing by the unsupervised spiking neural network.

17. The computer program product of claim 16 , wherein the first machine learning model is run on a neuromorphic chip.

18. The computer program product of claim 15 , wherein the first parameter set values are obtained by training the first machine learning model based on training input data and ground-truth label data for the training input data.

19. The computer program product of claim 15 , wherein the second parameter set values of the second machine learning model are obtained through a supervised learning process.

20. The computer program product of claim 15 , wherein the machine-executable instructions, when executed by the device, further cause the device to:

in accordance with an absence of the first ground-truth classification result for the first input data or the second ground-truth classification result for the second input data for a predetermined time period, discarding the first feature information or the second feature information.