2.5 Quintillion (2,500,000,000,000,000,000) bytes of data is generated every day i.e. close to 50,000 GB per second. This would fill about 10 million Blu Ray discs which when stacked up equal the height of 4 Eifel Towers. Today 3 billion people have access to the internet – this is equal to the population of 1960. However, a staggering 90% of this data generated is unstructured. Such big unstructured data thus provides an opportunity for data analytics and the development of artificial intelligence for such analyses.
Clarifai, a startup, is building software to help large organizations perform deep learning, a type of artificial intelligence which relies on computer models that simulate the way information is processed by the human brain. The process of deep learning basically involves the training of systems called artificial neural networks on information like pictures, videos, audio etc., which is then analyzed by the software which in response generates fresh information based around its predictions.
As of now, Clarifai’s algorithm basically helps users (largely businesses right now but it soon plans to roll out its services to individuals as well) search through a vast database of pictures and videos. On the flip side, the algorithm also scans and predicts what’s in a picture and video.
Every day there are 311 million Instagram posts, 17000 hours of YouTube footage uploaded, 300 million photos and 4.75 billion pieces of content uploaded and shared on Facebook. These statistics, from 2 companies and 3 platforms, alone present a colossal opportunity for a company like Clarifai, not taking into account the millions of other businesses worldwide and billions of individual users that could benefit from it. Clarifai could essentially use its platform with the embedded deep learning algorithm to target stock image companies with large databases of photos and videos, e-commerce companies such as Amazon and eBay, dating websites, video blogs, online streaming sites, cloud companies, architecture firms, educational institutions, gaming and visual effects companies, movie and animation studios as well as to targeted advertisers. The opportunities and the clientele are boundless.
While it could be debated that a lot of Clarifai’s existing and potential clients are big businesses that could essentially develop the technology for themselves, the challenge is that it takes years and experts in deep learning to get the technology right. It basically requires building learning models that process all the visual data and the difficulty with machine learning currently is that it develops, trains and perfects itself with use over time. Additionally, the field also presents challenges in terms of shortage of good talent. Most of the well-known names in this field have been snatched up by the few big players.
Though it may seem too advanced for a startup, Clarifai is currently in the process of developing an Application Program Interface or an API that will let software developers access Clarifai’s search technology over the web, essentially extending Clarifai’s reach and platform. With such advancement in Clarifai’s technology and its algorithm, Clarifai presents stiff competition to companies like Google and Facebook – companies’ whose future largely depends on the development of artificial intelligence and which have substantially large and advance projects lined up.
Such technology not only has massive prospects for big businesses but also represents the fundamentals on which the future of technology relies on. Deep learning has led to vast improvements in Android’s voice recognition system as well as helped Microsoft in creating a live voice translation system called Skype Translate.
It is of no surprise therefore that Clarifai has raised several rounds of funding worth millions of dollars with many big names betting on the success of its technology while being a lucrative company that many big tech firms are looking to snatch up.