Collection flow for confidential data
View Patent ↗In an example embodiment, an attribute interference model is trained by a machine learning algorithm to output missing attribute values from a member profile of a social networking service. In an attribute inference phase, an identification of a member of a social networking service is obtained. A member profile corresponding to the member of the social networking service is retrieved using the identification. The member profile is then passed to the attribute inference model to generate one or more missing attribute values for the member profile. A collection flow, defined in a user interface of a computing device, is modified based on the generated one or more missing attribute values, the collection flow defining a sequence of screens for collecting confidential data. The modified collection flow is then presented to the member in the user interface to collect confidential data from the member.
1. A computerized method comprising:
in a training phase:
obtaining a plurality of sample labeled member profiles with sample missing attribute values identifying attribute values missing from the sample labeled member profiles;
for each of the plurality of sample labeled member profiles:
extracting one or more features from the sample labeled member profile;
feeding the extracted one or more features and the sample missing attribute values into a machine learning algorithm to train an attribute inference model to output attribute values for the sample missing attribute values based on the extracted one or more features;
in an attribute inference phase:
obtaining an identification of a member of a social networking service
retrieving, using the identification, a member profile corresponding to the member of the social networking service;
passing the member profile to the attribute inference model to generate one or more attribute values for attribute values missing in the member profile;
modifying a collection flow, defined in a user interface of a computing device, based on the generated one or more missing attribute values, the collection flow defining a sequence of screens for collecting confidential data; and
presenting the modified collection flow to the member in the user interface to collect confidential data from the member.
2. The method of claim 1 , wherein the attribute inference model is further trained to output a confidence score for each of the one or more missing attribute values and the modifying includes comparing, for each of the one or more missing attribute values, the confidence score to at least one threshold and modifying the collection flow for missing attribute values having confidence scores transgressing the at least one threshold.
3. The method of claim 1 , wherein the modifying includes prepopulating the generated one or more missing attribute values in one or more screens of the user interface.
4. The method of claim 1 , wherein the one or more features includes co-occurrence of the missing one or more attribute values with attribute values contained in the sample member profiles.
5. The method of claim 1 , wherein the one or more attribute values include one or more skills.
6. The method of claim 1 , further comprising:
adding the confidential data to a first submission table;
adding at least one of the generated one or more missing attribute values to a second submission table along with one or more attribute values contained in the member profile; and
assigning the confidential data from the first submission table to a slice table corresponding to a slice including, from the second submission table, the one or more missing attribute values and the one or more attribute values contained in the member profile.
7. The method of claim 6 , wherein the first submission table is the same as the second submission table.
8. A system comprising:
a computer-readable medium having instructions stored there on, which, when executed by a processor, cause the system to:
in a training phase:
obtain a plurality of sample labeled member profiles with sample missing attribute values identifying attribute values missing from the sample labeled member profiles;
for each of the plurality of sample labeled member profiles:
extract one or more features from the sample labeled member profile;
feed the extracted one or more features and the sample missing attribute values into a machine learning algorithm to train an attribute inference model to output attribute values for the sample missing attribute values based on the extracted one or more features;
in an attribute inference phase:
obtain an identification of a member of a social networking service
retrieve, using the identification, a member profile corresponding to the member of the social networking service;
pass the member profile to the attribute inference model to generate one or more attribute values for attribute values missing in the member profile;
modify a collection flow, defined in a user interface of a computing device, based on the generated one or more missing attribute values, the collection flow defining a sequence of screens for collecting confidential data; and
present the modified collection flow to the member in the user interface to collect confidential data from the member.
9. The system of claim 8 , wherein the instructions further cause the system to:
add the confidential data to a first submission table;
add at least one of the generated one or more missing attribute values to a second submission table along with one or more attribute values contained in the member profile; and
assign the confidential data from the first submission table to a slice table corresponding to a slice including, from the second submission table, the one or more missing attribute values and the one or more attribute values contained in the member profile.
10. The system of claim 9 , wherein the first submission table is the same as the second submission table.
11. The system of claim 8 , wherein the attribute inference model is further trained to output a confidence score for each of the one or more missing attribute values and the modifying includes comparing, for each of the one or more missing attribute values, the confidence score to at least one threshold and modifying the collection flow for missing attribute values having confidence scores transgressing the at least one threshold.
12. The system of claim 8 , wherein the modifying includes prepopulating the generated one or more missing attribute values in one or more screens of the user interface.
13. The system of claim 8 , wherein the one or more features includes co-occurrence of the missing one or more attribute values with attribute values contained in the sample member profiles.
14. The system of claim 8 , wherein the one or more attribute values include one or more skills.
15. A non-transitory machine-readable storage medium comprising instructions, which when implemented by one or more machines, cause the one or more machines to perform operations comprising:
in a training phase:
obtaining a plurality of sample labeled member profiles with sample missing attribute values identifying attribute values missing from the sample labeled member profiles;
for each of the plurality of sample labeled member profiles:
extracting one or more features from the sample labeled member profile;
feeding the extracted one or more features and the sample missing attribute values into a machine learning algorithm to train an attribute inference model to output attribute values for the sample missing attribute values based on the extracted one or more features,
in an attribute inference phase:
obtaining an identification of a member of a social networking service
retrieving, using the identification, a member profile corresponding to the member of the social networking service;
passing the member profile to the attribute inference model to generate one or more attribute values for attribute values missing in the member profile;
modifying a collection flow, defined in a user interface of a computing device, based on the generated one or more missing attribute values, the collection flow defining a sequence of screens for collecting confidential data; and
presenting the modified collection flow to the member in the user interface to collect confidential data from the member.
16. The non-transitory machine-readable storage medium of claim 15 , wherein the attribute inference model is further trained to output a confidence score for each of the one or more missing attribute values and the modifying includes comparing, for each of the one or more missing attribute values, the confidence score to at least one threshold and modifying the collection flow for missing attribute values having confidence scores transgressing the at least one threshold.
17. The non-transitory machine-readable storage medium of claim 15 , wherein the modifying includes prepopulating the generated one or more missing attribute values in one or more screens of the user interface.
18. The non-transitory machine-readable storage medium of claim 15 , wherein the one or more features includes co-occurrence of the missing one or more attribute values with attribute values contained in the sample member profiles.
19. The non-transitory machine-readable storage medium of claim 15 , wherein the instructions further comprise:
adding the confidential data to a first submission table;
adding at least one of the generated one or more missing attribute values to a second submission table along with one or more attribute values contained in the member profile; and
assigning the confidential data from the first submission table to a slice table corresponding to a slice including, from the second submission table, the one or more missing attribute values and the one or more attribute values contained in the member profile.
20. The non-transitory machine-readable storage medium of claim 19 , wherein the first submission table is the same as the second submission table.