Abstract
A data processing system can generate associations between user interests and target users based on advertiser features to allow for specific targeting of users with content such as media or advertisements. The target users can be associated with user behaviors such as, for example, client device capabilities, user browsing behavior and actions, or a combination thereof. The advertiser features can be, for example, campaign or ad groups, landing pages, websites, or a combination thereof. The user interests can be keywords, URLs, applications, and locations, or a combination thereof. The user interests, target users, and advertiser features can each be represented by a deep neural network. The data processing system can train the deep neural network to match conversion events between target users and advertiser features through user interests. The data processing system can then retrieve embedding representations of specific campaigns based on advertiser features.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
"Entity Suggestions in Custom Audience", Technical Disclosure Commons, (October 29, 2020)
https://www.tdcommons.org/dpubs_series/3725