Israeli firm Jinni has developed a new content discovery engine which combines natural language understanding (NLU) with voice-activated search to enable users to tell their TV what kind of programme or film they wish to see.
The Jinni NLU engine leverages the company’s unique Entertainment Genome to interpret natural human speech and derive the underlying meaning to enable rich, intuitive interaction between users and their TVs.
It does this by comparing a user’s search request to an index of “genes” which have been assigned to thousands of programmes and films to describe mood, style, plot, setting and more.
New titles are automatically indexed in a database called the Movie Genome via analysis of user reviews and synopses, using a proprietary natural language processing (NLP) solution.
For example, a user can tell their TV: “Is there anything witty and romantic on TV tonight?” or “I want to watch something funny about an obnoxious boss” and the Jinni engine will search across linear TV, video-on-demand and any other available video catalogue.
“Consumer demand has changed dramatically and today people expect to be able to interact with technology in a very natural, personalised way,” explained Jinni co-founder and CEO, Yosi Glick.
“This is the core belief that inspired our semantic approach to video discovery and has allowed us to bring such an advanced NLU solution to market so quickly.”