21:24 18 October 2017
Researchers at EPFL’s Embedded Systems Laboratory (ESL) have found a way to reduce the power and storage needed to stream videos while enhancing user experience at the same time.
Marina Zapater Sancho, post-doctoral researcher at ESL, explains: "Platforms like YouTube or Netflix use two systems, both of which are inefficient. They store either one copy of a video in the highest-quality format possible, or dozens of copies in different formats. The former can result in slow and choppy streaming, while the latter takes up huge amounts of server storage and consequently eats up lots of power."
The researchers said that using a method they developed, the power requirement could be reduced by up to 20per cent. This can have a major impact because video streaming now makes up 80per cent of internet traffic worldwide.
The method involves machine learning. Arman Iranfar, a PhD student at ESL and co-author of the study explains: "Computers learn from experience. We exposed them to many different scenarios, such as 1,000 people playing a video, each from a different device. The computers remembered the series of actions that led to positive outcomes and reproduced them."
Using the machine learning method, applications that encode videos can learn to better allocate resources while optimising factors like power consumption, performance, quality and compression.
Disclaimer: Supanet is not responsible for, and disclaims any and all liability for the content of comments written by contributors to this website