How Media Technologies are Leveraging Machine Learning and AI
Artificial intelligence (AI) and machine learning are hot topics in technology, and news articles highlighting a new and exciting use case for them seem to spring up every day. AI refers to the ability of machines to demonstrate intelligence, while machine learning is a subset of AI that involves the use of statistical techniques that give computers the ability to learn and improve without being programmed.
The evolution and ubiquity of digital media are important distinguishing factors of the Information Age, and both image and video technologies continue to grow at a terrific pace, as evidenced by forecasted statistics. By 2021, for example, one million minutes of video content will cross global IP networks every second and video traffic will account for 82 percent of all consumer Internet traffic.
The evolution of video and image naturally intertwines with the growth of machine learning and AI. However, before discussing that further, let's take a look at some other relevant trends:
Digital asset management systems.
Most modern businesses understand the need for a web presence, rich in various digital media. However, maintaining visibility over a wealth of digital assets quickly becomes problematic, particularly for, say, an e-commerce business or an online publisher. Cloud-based digital asset management services have evolved to deal with this issue, and these digital asset management systems enable owners to tag, control access to, and easily edit digital assets.
Content creators understand that increasingly limited attention spans coupled with the slew of content that gets uploaded online daily necessitates bite-sized storytelling. Such trends are evident on platforms like Instagram Stories and Snapchat, where content creators and brands attempt to convey a message and impact viewers in short bursts of time, measured in seconds.
Many apps are now using facial recognition features to enhance security or improve user experience.
Related article: Emotionally Intelligent Design: Why You Need It in Your Mobile App
These trends are exciting, no doubt, but AI and machine learning have the potential for some incredible use cases in digital media technology. Below you'll find five ways in which media technologies can leverage AI and machine learning.
1. Intelligent video content recommendations
Traditional forms of media consumption such as cable television are being replaced by subscription streaming services - also known as subscription-video-on-demand - such as Netflix, Amazon Prime, and more. Video sharing platforms like Youtube also continue to grow.
These services rely quite heavily on intelligent content recommendations to keep viewers always wanting more, and this is an area in which machine learning is becoming prominent. Youtube, for example, uses machine learning models and algorithms to drive 70 percent of what people watch on the platform. Netflix also uses AI to suggest shows worth bingeing on to subscribers.
2. Personalized ads
Many digital content creators and publishing platforms rely on advertising as a means of income, and the rise of video means that the digital advertising industry will continue to grow. Displaying the same advertisement to all viewers of a video no longer suffices, and machine learning algorithms are beginning to influence this space, with targeted and personalized video ads.
Because machine learning caters for continuous optimization over time, such models have the ability to learn from individual consumer responses to videos and provide personalized ads based on a number of other factors, including viewer geography and demographics.
Related article: The Life Cycle of Online Ads: from User Preferences to Ad Revenue
3. Processing low-light images
In what is a really exciting use of machine learning's capabilities for processing images, a team at the University of Illinois Urbana-Champaign managed to use machine learning to dramatically improve images shot in low light.
The machine learning pipeline can almost flawlessly correct photos shot in near darkness, which could revolutionize low-light photography and lead to new possibilities for websites, publishers, and content creators in terms of the stories they can tell through their images.
4. Color-matching videos
Any content creator knows that in between recording a video and publishing it on sites like Youtube is an often painful editing process that takes raw clips and forms a smooth video which tells a story.
One of the most frustrating parts of editing is trying to match colors between videos shot using multiple cameras. Creators often use smartphones, DSLRs, and drones in the same video to give different perspectives on a scene and better engage audiences. Adobe is looking to reduce editing headaches with AI-driven color-match video editing. Adobe Sensei provides intelligent color-matching capabilities through its AI and machine learning platform.
5. Reconstructing damaged images
In April 2018, several outlets reported on technology company Nvidia's use of AI to reconstruct damaged images. The process, referred to as "image inpainting", uses a deep learning model to intelligently handle missing aspects of any shape, size location, or distance from the image borders and replaces them with realistic computer-generated alternatives.
As this technology evolves, it could be used to even restore old black and white photos for publishing online, or, at a minimum, to allow media publishers and businesses to derive more use from all their media assets, even those that were previously unusable due to damage.
Even without AI, video and image technologies continue to improve thanks to cloud digital asset management systems and other innovations. However, the range of uses of AI and machine learning for digital media promises unheard of advances in the way people create and consume both images and videos.
This is a guest post by Limor Wainstein. Limor is a technical writer and editor at Agile SEO, a boutique digital marketing agency focused on technology and SaaS markets. She has over 10 years' experience writing technical articles and documentation for various audiences, including technical on-site content, software documentation, and dev guides. She specializes in big data analytics, computer/network security, middleware, software development and APIs. Follow her on LinkedIn and Twitter.