Electronic message classification and delivery using a neural network architecture

1. A system for electronic message classification and delivery using a neural network architecture, comprising:

one or more computing devices associated with one or more users; and

at least one computer processing system in communication with the one or more computing devices over at least one network, the at least one computer processing system including:

a memory device for storing program code; and

at least one processor operatively coupled to the memory device and configured to execute program code stored on the memory device to:

receive one or more inputs associated with one or more e-mails corresponding to the one or more users across the at least one network;

classify the one or more e-mails by performing natural language processing based on one or more sets of input-aware filters conditioned on respective ones of the one or more inputs, said input-aware filters generated by an adaptive convolutional filter generation framework that uses a meta-network to generate the input-aware filters, said one or more sets of input-aware filters conditioned on an input sentence and adaptively applied to text sequences such that the one or more sets of input-aware filters are variable from input sentence-to-sentence; and

permit the one or more users access to the one or more classified e-mails via the one or more computing devices.

2. The system of claim 1 , wherein the at least one processor is further configured to execute program code stored on the memory device to generate the one or more sets of filters by implementing one or more encoders to encode the one or more inputs into one or more respective vectors, and one or more decoders to determine the one or more sets of filters based on the one or more vectors.

3. The system of claim 2 , wherein the at least one processor is configured to classify the one or more e-mails by performing at least one operation selected from the group consisting of: determining whether the one or more e-mails are unsolicited, assigning the one or more e-mails to one or more folders, and assigning one or more tags to the one or more e-mails.

4. The system of claim 2 , wherein the one or more encoders each include a convolutional layer, and the one more decoders each include a deconvolutional layer.

5. The system of claim 1 , wherein the at least one processor is further configured to execute program code stored on the memory device to perform adaptive convolution by applying the one or more sets of filters to respective ones of the one or more inputs to generate one or more representations.

6. The system of claim 5 , wherein the at least one processor is configured to apply the one or more sets of filters by applying one or more convolutional layers.

7. The system of claim 5 , wherein the at least one processor is configured to classify the one or more e-mails by performing the natural language process task based on the one or more sets of filters by:

predicting one or more probabilities over one or more respective classes based on the one or more representations; and

outputting one or more labels based on the one or more probabilities.

8. A computer-implemented method for electronic message classification and delivery using a neural network architecture, comprising:

receiving, by at least one computer processing system including at least one processor operatively coupled to a memory, one or more inputs associated with one or more e-mails corresponding to one or more users across at least one network;

classifying, by the at least one computer processing system, the one or more e-mails by performing natural language processing based on one or more sets of input-aware filters conditioned on respective ones of the one or more inputs, said input-aware filters generated by an adaptive convolutional filter generation framework that uses a meta-network to generate the input-aware filters, said one or more sets of input-aware filters conditioned on an input sentence and adaptively applied to text sequences such that the one or more sets of input-aware filters are variable from input sentence-to-sentence; and

permitting, by the at least one computer processing system, the one or more users access to the one or more classified e-mails via the one or more computing devices.

9. The method of claim 8 , wherein classifying the one or more e-mails includes performing at least one operation selected from the group consisting of: determining whether the one or more e-mails are unsolicited, assigning the one or more e-mails to one or more folders, and assigning one or more tags to the one or more e-mails.

10. The method of claim 8 , further comprising generating, by the at least one computer processing system, the one or more sets of filters by implementing one or more encoders to encode the one or more inputs into one or more respective vectors, and one or more decoders to determine the one or more sets of filters based on the one or more vectors.

11. The method of claim 10 , wherein the one or more encoders each include a convolutional layer, and the one more decoders each include a deconvolutional layer.

12. The method of claim 8 , further comprising performing, by the at least one computer processing system, adaptive convolution by applying the one or more sets of filters to respective ones of the one or more inputs to generate one or more representations.

13. The method of claim 12 , wherein applying the one or more sets of filters further includes applying one or more convolutional layers.

14. The method of claim 12 , wherein classifying the one or more e-mails by performing the natural language process task based on the one or more sets of filters further includes:

predicting one or more probabilities over one or more respective classes based on the one or more representations; and

outputting one or more labels based on the one or more probabilities.

15. A computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method for electronic message classification and delivery using a neural network architecture, the method performed by the computer comprising:

receiving one or more inputs associated with one or more e-mails corresponding to one or more users across at least one network;

classifying the one or more e-mails by performing natural language processing based on one or more sets of input-aware filters conditioned on respective ones of the one or more inputs, said input-aware filters generated by an adaptive convolutional filter generation framework that uses a meta-network to generate the input-aware filters, said one or more sets of input-aware filters conditioned on an input sentence and adaptively applied to text sequences such that the one or more sets of input-aware filters are variable from input sentence-to-sentence; and

permitting the one or more users access to the one or more classified e-mails via the one or more computing devices.

16. The computer program product of claim 15 , wherein classifying the one or more e-mails includes performing at least one operation selected from the group consisting of: determining whether the one or more e-mails are unsolicited, assigning the one or more e-mails to one or more folders, and assigning one or more tags to the one or more e-mails.

17. The computer program product of claim 15 , wherein the method further comprises generating the one or more sets of filters by implementing one or more encoders to encode the one or more inputs into one or more respective vectors, and one or more decoders to determine the one or more sets of filters based on the one or more vectors.

18. The computer program product of claim 17 , wherein the one or more encoders each include a convolutional layer, and the one more decoders each include a deconvolutional layer.

19. The computer program product of claim 15 , wherein the method further comprises performing adaptive convolution by applying the one or more sets of filters to respective ones of the one or more inputs to generate one or more representations.

20. The computer program product of claim 19 , wherein classifying the one or more e-mails by performing the natural language process task based on the one or more sets of filters further includes:

predicting one or more probabilities over one or more respective classes based on the one or more representations; and

outputting one or more labels based on the one or more probabilities.