Creating structured data to drive search, recommendations and discovery with the help of machine learning extracts and semantic graphs
Today’s consumers have an ocean of content, including movies, programs, and news and short-form videos to choose from an array of linear and streaming services. However, the abundance of choice can be overwhelming, as viewers cannot find what they want to watch quickly and easily. The primary reason for this frustration is the lack of structured data as it drives search, recommendations and discovery. In their peer-reviewed technical paper, Lijin Chungapalli, Senior Software Engineer, Metadata Engineering Group, and Venkata Babji Perambattu, Engineering Manager at TiVo, demonstrate how to overcome this challenge.
Effective entertainment discovery solutions require a deeper understanding of content, and one approach to harnessing this knowledge is extracting semantically relevant metadata. In their paper, Machine learning extracts and semantic graphs: Creating structured data to drive search, recommendations and discovery, the authors explain how to do just that.
Historically, semantic graphs have helped a great deal in question-answering (Dali et al.), and text summarization (Moawrd and Ared.) Furthermore, a combination of semantic graphs and machine learning can be used to automatically generate structured data, recognize important entities/keywords and create weighted connections.
Therefore, in this paper, the authors delve into ways to leverage the nodes in a semantic graph. According to the authors, the nodes can be used to train a machine-learning model that will automatically determine the relevance of an entity in a given blurb of text. This will help generate more relevant search results and recommendations.
By inferring relevant entities through these underlying technologies, metadata results are richer and more meaningful, enabling faster decision making for the viewer and stronger viewership for the content owner.
Click here to download the full technical paper.