Optimized graph traversal
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for optimized graph traversal are disclosed. In one aspect, a method includes the actions of receiving a given phrase that is input through a user interface by a digital component provider. The actions further include determining an entity that is being referred to by the given phrase. The actions further include identifying properties of the entity. The actions further include selecting a subset of the properties that were identified for the entity. The actions further include identifying additional phrases. The actions further include updating the user interface to present at least some of the additional phrases with programmatic controls that assign one or more of the additional phrase as distribution criteria for digital components of the digital component provider in response to activation of the programmatic controls.
1. A system comprising:
one or more memory devices storing computer executable instructions; and
one or more processors configured to execute the instructions and perform operations comprising:
receiving, through a user interface, a phrase submitted by a particular digital component provider;
providing the phrase as an input to a selection model trained based on previous phrases submitted by other digital component providers and corresponding content specified by the other digital component providers;
receiving, from the selection model and in response to the phrase being provided as the input to the selection model, a group of phrases that are related to the phrase, wherein the group of phrases omits one or more related phrases that have a lower performance characteristic than the phrase submitted by the particular digital component provider; and
updating the user interface to present at least a portion of the group of phrases with programmatic controls that assign one or more phrases of the group of phrases as distribution criteria that control distribution of digital components provided by the particular digital component provider in response to interaction with the programmatic controls by the particular digital component provider.
2. The system of claim 1 , wherein the one or more processors are configured to train the selection model based on the previous phrases and related entities identified within a knowledge graph using the previous phrases, wherein the one or more processors are configured to identify the related entities that are related to each previous phrase by traversing a portion of the knowledge graph.
3. The system of claim 2 , wherein the related entities used to train the selection model are within a threshold distance of a node corresponding to entities referred to by each previous phrase.
4. The system of claim 1 , wherein the group of phrases received from the selection model comprise a minimum spanning tree among the phrase.
5. The system of claim 1 , wherein the one or more processors are configured to perform operations comprising:
receiving an activation of a particular programmatic control of the programmatic controls, the particular programmatic control assigning a particular phrase of the group of phrases as distribution criteria for a particular digital component of the digital components of the particular digital component provider; and
updating the selection model based on the activation of the particular programmatic control assigning the particular phrase of the group of phrases as distribution criteria for the particular digital component of the digital components of the particular digital component provider.
6. The system of claim 1 , wherein the selection model is configured to reduce a number of accesses to a knowledge graph when identifying the group of phrases.
7. The system of claim 1 , wherein the one or more processors are configured to perform operations comprising identifying entities that are related to the phrase and the group of phrases that are related to the entities.
8. A computer-implemented method comprising:
receiving, through a user interface, a phrase submitted by a particular digital component provider;
providing the phrase as an input to a selection model trained based on previous phrases submitted by other digital component providers and corresponding content specified by the other digital component providers;
receiving, from the selection model and in response to the phrase being provided as the input to the selection model, a group of phrases that are related to the phrase, wherein the group of phrases omits one or more related phrases that have a lower performance characteristic than the phrase submitted by the particular digital component provider; and
updating the user interface to present at least a portion of the group of phrases with programmatic controls that assign one or more phrases of the group of phrases as distribution criteria that control distribution of digital components provided by the particular digital component provider in response to interaction with the programmatic controls by the particular digital component provider.
9. The method of claim 8 , further comprising:
training the selection model based on the previous phrases and related entities identified within a knowledge graph using the previous phrases; and
identifying the related entities that are related to each previous phrase by traversing a portion of the knowledge graph.
10. The method of claim 9 , wherein the related entities used to train the selection model are within a threshold distance of a node corresponding to entities referred to by each previous phrase.
11. The method of claim 8 , wherein the group of phrases received from the selection model comprise a minimum spanning tree among the phrase.
12. The method of claim 8 , further comprising:
receiving an activation of a particular programmatic control of the programmatic controls, the particular programmatic control assigning a particular phrase of the group of phrases as distribution criteria for a particular digital component of the digital components of the particular digital component provider; and
updating the selection model based on the activation of the particular programmatic control assigning the particular phrase of the group of phrases as distribution criteria for the particular digital component of the digital components of the particular digital component provider.
13. The method of claim 8 , wherein the selection model is configured to reduce a number of accesses to a knowledge graph when identifying the group of phrases.
14. The method of claim 8 , further comprising identifying entities that are related to the phrase and the group of phrases that are related to the entities.
15. A non-transitory computer-readable medium storing instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:
receiving, through a user interface, a phrase submitted by a particular digital component provider;
providing the phrase as an input to a selection model trained based on previous phrases submitted by other digital component providers and corresponding content specified by the other digital component providers;
receiving, from the selection model and in response to the phrase being provided as the input to the selection model, a group of phrases that are related to the phrase, wherein the group of phrases omits one or more related phrases that have a lower performance characteristic than the phrase submitted by the particular digital component provider; and
updating the user interface to present at least a portion of the group of phrases with programmatic controls that assign one or more phrases of the group of phrases as distribution criteria that control distribution of digital components provided by the particular digital component provider in response to interaction with the programmatic controls by the particular digital component provider.
16. The computer-readable medium of claim 15 , wherein the instructions cause the one or more computers to perform operations further comprising:
training the selection model based on the previous phrases and related entities identified within a knowledge graph using the previous phrases; and
identifying the related entities that are related to each previous phrase by traversing a portion of the knowledge graph.
17. The computer-readable medium of claim 16 , wherein the related entities used to train the selection model are within a threshold distance of a node corresponding to entities referred to by each previous phrase.
18. The computer-readable medium of claim 15 , wherein the group of phrases received from the selection model comprise a minimum spanning tree among the phrase.
19. The computer-readable medium of claim 15 , wherein the instructions cause the one or more computers to perform operations further comprising:
receiving an activation of a particular programmatic control of the programmatic controls, the particular programmatic control assigning a particular phrase of the group of phrases as distribution criteria for a particular digital component of the digital components of the particular digital component provider; and
updating the selection model based on the activation of the particular programmatic control assigning the particular phrase of the group of phrases as distribution criteria for the particular digital component of the digital components of the particular digital component provider.
20. The computer-readable medium of claim 15 , wherein the selection model is configured to reduce a number of accesses to a knowledge graph when identifying the group of phrases.