Native advertisements mimic the look and feel of a publisher’s content slots and are used to monetize content inventories. Currently, native ad styles are based on hand-written rules and heuristic which can result in ad styles that are not consistently high-quality. This disclosure describes the use of machine learning models to automatically generate native ad styles for online or mobile publishers. For a given content slot and elements therein, native ad styles are automatically generated that closely resemble publisher style, match page context, and perform well. Local and global features of the content slot and elements are used to generate the native ad. The techniques can automatically incorporate new abstractions.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Lin, Bo, "Automatic generation of native ad styles", Technical Disclosure Commons, (March 30, 2018)