Pattern-based classification
A method includes receiving interaction data that indicates, for each given interaction among multiple interactions that occurred at a client device, (i) an event type an (ii) a delay period specifying an amount of time between the given event and a previous event that occurred prior to the given event, encoding each given interaction into an encoded interaction having a standardized format that is a combination of (i) the event type and (ii) the delay period, generating an interaction signature that includes sequence of encoded interactions, processing the sequence of encoded interactions using a model trained to label sequences of user interactions as valid or invalid, including labelling, using the model, a sequence of encoded interactions as invalid, and preventing distribution of a set of content to an entity that performed the sequence of encoded interactions in response to a subsequently identified request to provide content to the entity.
1. A method performed by one or more data processing apparatus, the method comprising:
receiving multiple sets of interaction data for a given entity, each set of interaction data indicating, for each given interaction among multiple interactions that occurred at a client device associated with the given entity, (i) an event type and (ii) a delay period specifying an amount of time between the given interaction and a previous interaction that occurred prior to the given interaction;
encoding the interaction data for each given interaction into encoded interaction data having a standardized format that is a combination of (i) the event type of the given interaction and (ii) the delay period specified by the interaction data for the given interaction;
providing the encoded interaction data for each user interaction in a sequence of user interactions to a trained machine learning model that is trained to classify sequences of user interactions as valid or invalid;
receiving, as an output of the trained machine learning model, a classification of the sequence of user interactions; and
distributing content to the given entity based on the classification of the sequence of user interactions.
2. The method of claim 1 , wherein distributing content to the given entity based on the classification of the sequence of user interactions comprises distributing content to the given entity in response to the classification of the sequence of user interactions being valid.
3. The method of claim 1 , wherein distributing content to the given entity based on the classification of the sequence of user interactions comprises reducing an amount of content distributed to the given entity in response to the classification of the sequence of user interactions being invalid.
4. The method of claim 1 , comprising preventing distribution of content to entities for which a classification output by the trained machine learning model is invalid.
5. The method of claim 1 , wherein the trained machine learning model comprises a deep neural network.
6. The method of claim 1 , comprising adjusting distribution criteria for a digital component based on at least on the classification of the sequence of user interactions.
7. The method of claim 6 , wherein adjusting the distribution criteria for the digital component comprises adjusting the distribution criteria based on classifications for multiple entities.
8. A system comprising:
one or more processors; and
one or more memory elements including instructions that, when executed, cause the one or more processors to perform operations including:
receiving multiple sets of interaction data for a given entity, each set of interaction data indicating, for each given interaction among multiple interactions that occurred at a client device associated with the given entity, (i) an event type and (ii) a delay period specifying an amount of time between the given interaction and a previous interaction that occurred prior to the given interaction;
encoding the interaction data for each given interaction into encoded interaction data having a standardized format that is a combination of (i) the event type of the given interaction and (ii) the delay period specified by the interaction data for the given interaction;
providing the encoded interaction data for each user interaction in a sequence of user interactions to a trained machine learning model that is trained to classify sequences of user interactions as valid or invalid;
receiving, as an output of the trained machine learning model, a classification of the sequence of user interactions; and
distributing content to the given entity based on the classification of the sequence of user interactions.
9. The system of claim 8 , wherein distributing content to the given entity based on the classification of the sequence of user interactions comprises distributing content to the given entity in response to the classification of the sequence of user interactions being valid.
10. The system of claim 8 , wherein distributing content to the given entity based on the classification of the sequence of user interactions comprises reducing an amount of content distributed to the given entity in response to the classification of the sequence of user interactions being invalid.
11. The system of claim 8 , wherein the operations comprise preventing distribution of content to entities for which a classification output by the trained machine learning model is invalid.
12. The system of claim 8 , wherein the trained machine learning model comprises a deep neural network.
13. The system of claim 8 , wherein the operations comprise adjusting distribution criteria for a digital component based on at least on the classification of the sequence of user interactions.
14. The system of claim 13 , wherein adjusting the distribution criteria for the digital component comprises adjusting the distribution criteria based on classifications for multiple entities.
15. A non-transitory computer storage medium encoded with instructions that when executed by a computing system cause the computing system to perform operations comprising:
receiving multiple sets of interaction data for a given entity, each set of interaction data indicating, for each given interaction among multiple interactions that occurred at a client device associated with the given entity, (i) an event type and (ii) a delay period specifying an amount of time between the given interaction and a previous interaction that occurred prior to the given interaction;
encoding the interaction data for each given interaction into encoded interaction data having a standardized format that is a combination of (i) the event type of the given interaction and (ii) the delay period specified by the interaction data for the given interaction;
providing the encoded interaction data for each user interaction in a sequence of user interactions to a trained machine learning model that is trained to classify sequences of user interactions as valid or invalid;
receiving, as an output of the trained machine learning model, a classification of the sequence of user interactions; and
distributing content to the given entity based on the classification of the sequence of user interactions.
16. The non-transitory computer storage medium of claim 15 , wherein distributing content to the given entity based on the classification of the sequence of user interactions comprises distributing content to the given entity in response to the classification of the sequence of user interactions being valid.
17. The non-transitory computer storage medium of claim 15 , wherein distributing content to the given entity based on the classification of the sequence of user interactions comprises reducing an amount of content distributed to the given entity in response to the classification of the sequence of user interactions being invalid.
18. The non-transitory computer storage medium of claim 15 , wherein the operations comprise preventing distribution of content to entities for which a classification output by the trained machine learning model is invalid.
19. The non-transitory computer storage medium of claim 15 , wherein the trained machine learning model comprises a deep neural network.
20. The non-transitory computer storage medium of claim 15 , wherein the operations comprise adjusting distribution criteria for a digital component based on at least on the classification of the sequence of user interactions.