Thank Stanford researchers for Puffer, a free and open source live TV streaming service that uses AI to improve video-streaming algorithms

2 min read

A team of researchers from Standford University launched a new, free, and open source TV streaming service called Puffer, as part of their non-profit academic research study. It is led by Francis Yan, a doctoral student, Computer Science, Stanford University, Sadjad Fouladi, Hudson Ayers, and Chenzhi Zhu.

Puffer uses machine learning to improve video-streaming algorithms. “We are trying to figure out how to teach a computer to design new algorithms that reduce glitches and stalls in streaming video (especially over wireless networks and those with limited capacities, such as in rural areas),” say the researchers.

Puffer is mainly focused on three algorithms, namely, “congestion-control” (decides when to send each piece of data), “throughput forecasters” (predicts how long it will take to send a certain amount of data over an Internet connection), and “adaptive-bitrate” (ABR) (algorithms that decide what quality of video to send for best picture quality).

The project is limited to only 500 participants at a time. Participants would need to watch TV channels on Puffer and stream them over their Internet connections a few hours each week. As the participants are streaming the TV channels on the Puffer website, it will begin to automatically experiment with different algorithms to control the timing and quality of video sent to them. They will then analyze how the resulting computer-designed algorithm performs and work.

Puffer is a free service and doesn’t show any ads. Puffer is capable of only re-transmitting the free over-the-air broadcast TV signals and allows streaming of up to six TV stations. These include CBS (KPIX 5), NBC (KNTV 11), ABC (KGO 7), FOX (KTVU 2), PBS (KQED 9), and Univision (KDTV 14).

The  Puffer project has received funding in part by the NSF and DARPA. It has also received support from Google, Huawei, VMware, Dropbox, Facebook, and the Stanford Platform Lab.

“Puffer is unique from previous academic studies…we hope that this approach will produce substantial benefits over prior work, but only time will tell”, say the researchers.

For more information on Puffer, check out its official website.

Read Next

Researchers introduce a machine learning model where the learning cannot be proved

Researchers release unCaptcha2, a tool that uses Google’s speech-to-text API to bypass the reCAPTCHA audio challenge

Researchers design ‘AnonPrint’ for safer QR-code mobile payment: ACSC 2018 Conference