About the Patent Application
Facebook’s patent discloses a method of leveraging machine learning models to recommend media content based on user media interaction. The machine learning models analyze different features of audio and visual media to create recommendations for the user tailored to their specific interests.
The first machine learning model determines a possible video embedding site within video content. The video content can be analyzed on at least one of the following: a concept, an object, or a visual characteristic identified in the video content item. The second machine learning model determines a possible set of music embeddings from music content. The music content can be analyzed based on at least one of the following features: a title, a lyric, a genre, or a spectrogram associated with the set of music content items.
why This Patent Application Is interesting
Facebook is disclosing a way to recommend further audio and or video content from media the user has interacted with. This culminates into a ranked list that the user can interact with to discover further media recommendations.
This is the first patent in a new family for Facebook, so continuations of this application with more details as to how the system both works and is implemented are to be expected in the future.
One implementation of this patent application could be for automatically recommending music or video segments to be inserted into a user’s media stream. This could be as simple as adding a song to a play queue, or analyze a user’s audio-visual post and recommend music to go along with the video and additional video segments to insert into the post.
Machine Learning is no new technology for Facebook. The company’s research team notes, “Machine learning and Applied Machine Learning is essential to Facebook. It helps people discover new content and connect with the stories they care the most about.”
To learn more about the AI and engineering team at Facebook, visit their Facebook engineering page.
Written by John DeStefano, Technical Advisor
and Lauren Hawksworth, Marketing Administrator
March 16, 2021