At the F8 conference, Facebook CEO Mark Zuckerberg revealed how machine learning is being used to improve user experience and offer new marketing opportunities to the millions of businesses who use the site to advertise their services.
Facebook’s Applied Machine Learning team is responsible for its innovative AI functionality and hopes their advancements will set a new standard for how people and computers interact online. It’s a smart move on Facebook’s part, and gives them a timely edge over competitors. The global artificial intelligence market is expected to grow in value from USD 419.7 million in 2014 to USD 5.05 billion 2020, a market Facebook could become leaders in should their AI succeed in its endeavours.
Facebook announced plans to transform its Messenger platform into an e-commerce hub at last year’s F8, and this year revealed new partnerships and how this platform will interact with users. Uber, Spotify, KLM and Dropbox have all signed up to take advantage of the new system, which will use AI-powered chatbots to engage users. Facebook’s AI chatbots will support businesses of all kinds, providing a wide range of in-depth services such as marketing, delivery arrangement and customer service. The global market for social media messaging platforms is expected to be worth USD 27.2 billion by 2020, and Facebook will likely dominate this market thanks to their extremely popular Messenger service (Messenger is currently the second most popular app on iOS).
Talking the Talk
One way Facebook is using artificial intelligence to improve user experience and build a more global Facebook community is through its AI-powered automatic translation system. The system translates and posts content in the news feed for 800 million users on a monthly basis. It also allows users from different countries to communicate with one another regardless of the languages they speak. The AI-powered system has an advantage over traditional translation services in that it uses a causal language system featuring the nuances found in everyday language as opposed to a strictly general language structure. The result is a more natural translation free from the errors typically encountered when using direct translation systems. The system could give Facebook a strong entry point to the natural language processing market, one which is forecast to be worth USD 2.1 billion by 2024.
Finding the Right Photo
Facebook’s existing AI-powered automatic image classifiers help users find photos and specific information such as when and where a photo was taken. The Applied Machine Learning team want to advance this tech by analyzing how images are classified and what further segmentation can be introduced to image categories. The goal is to allow users to search for a specific image using image content instead of tags or accompanying text, as the feature will be able to recognize individual objects within an image. In addition to this, the team plan on adding an assistance feature for visually impaired people that helps with image searching and provides an audio description of the photo’s content. The image recognition market is expected to be worth USD 29.98 billion by 2020, suggesting heavy investment in this area.
Savvy Video Classification
As with its AI-powered automatic image classifiers, Facebook’s real-time video classification system will classify live videos instantaneously in real time, using computer vision technology to open and categorize videos without the need for tags or accompanying content. The team believes that the system could, in the future, support automatically generated closed captioning or use facial recognition to automatically tag people when they appear in a video.
Looking to the Future
Of course, the Applied Machine Learning team have plenty of ideas as to how they want Facebook’s AI technology to develop, and how they ultimately envision its AI to behave. Facebook’s long-term goal is to replicate the memory capabilities of humans and transfer this ability to AI, giving machines the ability to self-teach. Facebook believe this to be the key to unlocking the true potential of machine learning and artificial intelligence, but for now it seems the social network is content with using AI to better the online experiences of its users.
Facebook’s F8 conference featured a range of exciting new technological developments for the company, but the coverage on new artificial intelligence systems was by far the most interesting. From chatbots for businesses to real-time video classification, Facebook’s Applied Machine Learning team understand the potential of AI, and how this technology will benefit users and businesses alike.