Getting the film you want, when you want it

The Jinni genome for movies breaks down films into their ‘genetic’ traits, creating a state-of-the-art classification system that helps you choose what you want to see.

 Getting the film you want, when you want it


The team at Jinni aim to help viewers find and watch the movies they most want to see.

By David Halevi

Life’s too short to waste in front of the TV – that is, if you’re watching something that you aren’t fully enjoying. But with the plethora of entertainment choices available, zeroing-in on the activity or film that perfectly suits your mood has become more difficult than ever. It can make you feel stressed-out about your leisure time.

Israel’s Jinni has a solution to that leisure-time dilemma. The company has developed a system that analyzes a film’s ‘genetics,’ breaking down its ‘character traits’ and ‘features,’ so that "you get the film you want, when you want, without guesswork," says Jinni President and founder Yosi Glick.

Glick believes that the science of genetics is a natural match for the movies. Many different attributes govern mood and taste, he says – determining whether you’re in the mood for a comedy, drama, or even a "dramedy."

According to Glick, like genes in the human body, "movies have hundreds, if not thousands, of attributes, and the Jinni system defines those attributes and allows users to search through a huge database of movies, allowing them to precisely hone in on the movie that most fits their mood."

Analysis based on 2,200 parameters

However, determining these attributes and matching them to mood and desire is far more complicated than just labeling a movie a comedy, drama, action film, etc. The Jinni genome goes much further, taking into account almost every attribute that could apply to a film.

For example, says Glick, a film could be labeled a comedy, but within that category it could be slapstick, sophisticated, family-oriented, ‘mature situations,’ or one of dozens of other possibilities.

Add to that other criteria that may appeal to you, like location – a New York story, a California casual film, a Continental adventure – and whether you want something with or without violence, with specific actors, and so on, and you have an idea of the power of the Jinni engine.

"Jinni analyzes information about films, analyzing online reviews, viewer opinion, the plot itself, and many other sources, including ratings by Jinni users," Glick relates. Then, using its unique, patented textual analysis tools, Jinni’s engine "breaks the information down into the ‘genetic traits’ being described."

 Getting the film you want, when you want it

The Jinni system analyzes the genetics of a film to break down its character traits and features.

Jinni categorizes video content based on an impressive 2,200 different parameters. The information is stored in a database which users can search by attributes – so if they’re in the mood for an action film that deals with the relationship between children and parents that features chase scenes, and is funny as well – that’s exactly what they get.

Film genome has 80 percent accuracy

Glick cites Jinni as the most extensive and ambitious way for users to discover films. "Google and nearly all other search engines index by keyword, but Jinni indexes by meaning, so the system is much more useful for making decisions that involve emotions, like film preference."

The information being mined by Jinni is being used to build a ‘film genome,’ which, Glick says, will be freely available to anyone searching for something to watch. Without revealing specific numbers, Glick says that the genome is "well over 80 percent accurate," meaning that the vast majority of viewers walk away with exactly what they’re looking for.

Jinni has forged partnerships with a number of entertainment companies. For example, you can connect your Jinni account to your Netflix account, and search for films on the Netflix service according to Jinni database ‘genes.’ Far more ambitious, though, is the alliance between Jinni and Google on the latter’s new TV service, which will allow users to watch web videos and online programs on TV sets.

With the opening of the TV set to the web, viewers will have the option of watching literally tens of thousands of channels and programs, instead of just the dozens available on cable and satellite TV. That’s a lot of content to parse through, but Jinni is up to the job – and a future application, for the Web and for the currently-under-development Google TV platform based on the Android operating system, will enable viewers to quickly and efficiently find the web TV content they’re in the mood for.

From Google TV to the perfect coffee shop

Glick developed Jinni as a natural extension of his previous work in delivering content to computers – a field he worked in for nearly a decade. "I was actually one of the pioneers of the industry," he says. "We concentrated on delivering the content – compression, universal delivery to all computers, etc. – but search was never a priority, and as a result, we were stuck with the same single-dimension text-based search used for all Web searches."

With Jinni he felt he could "get into the mind of the writer and producer," classifying the attributes that s/he had in mind when they made the film. And the millions of users of the site and those on Netflix testify to the need for the service, which he says is the only one of its kind.

The Tel Aviv-based company, with about 20 employees, is a member of the Startup Factory and DFJ Tamir Fishman Ventures, which have funded Jinni with over $3 million in capital (Moshe Levin, a general partner at DFJ Tamir Fishman, is also chairman of Jinni).

While the company is concentrating on movies, the technology – and the construction of a genome – could apply to many more fields, like games, books, or lifestyle. For example, says Glick, the technology could be applied to find a coffee shop with easy parking suitable for a business meeting and plenty of power outlets for laptops.

But for now, Glick says, Jinni will concentrate on entertainment, focusing on enabling computers to ‘think’ like a human and understand mood, sentiment and meaning.