MIT and Netflix Testing AI-Based Algorithms to Curb Buffering

Waiting for a video to buffer may become an annoyance of the past. Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) are working on streaming algorithms that use AI to improve load rates and, thus, reduce buffering. Dubbed Pensieve, the new technology relies on machine learning to navigate the often-chaotic and ever-changing conditions of networks in real-time, based on a system of rewards (when the video loads smoothly) and penalties (when it’s interrupted). Meanwhile, Netflix is working on its own AI solution to address buffering.

VentureBeat reports that, “the system improves video streaming by 10 percent to 30 percent, and viewers rated its results 10 percent to 25 percent higher in overall quality over current algorithms.”


To rigorously test Pensieve, researchers “sent its AI-powered algorithm manager through a gauntlet of challenging real-world scenarios and purposefully exposed it to previously unknown network conditions to test its adaptive capabilities.” Pensieve passed every test, “maintaining the same video resolution as the best traditional streaming algorithms but with less rebuffering.”

“This sort of stress test shows that it can generalize well for new scenarios out in the real world,” said Hongzi Mao, a member of the MIT team and author of a related paper.

MIT’s Pensieve is “a first effort,” but not the only one attempting to solve the buffering issue. Earlier this year, Netflix debuted its own AI-powered algorithm called Dynamic Optimizer. “We’re allergic to rebuffering,” said Netflix vice president of product innovation Todd Yellin.

Its solution “analyzes video frame-by-frame in real time and selectively compresses each scene for higher image quality on slow connections,” and is also “smart enough to differentiate types of video content.” That means that, “action-packed sequences from the latest superhero flicks get an increase in bitrate, while simpler animated content is eased back.”

Netflix is also relying on deep learning and simulated neural networks to fine-tune its understanding of viewer preferences, thus “extending the average binge session.” Improved AI-powered streaming management also makes room for the bitrate requirements of 4K and virtual reality.