I am a Senior Researcher
in the
Mobility and Networking Research Group at
Microsoft Research Redmond
and the
Office of the CTO, Azure for Operators.
My research focuses on practical machine learning (ML) for networking,
seeking to create ML algorithms that are deployable on real-world networked systems
and build platforms for training and evaluating novel algorithms.
My research interests also span video streaming/conferencing/analytics,
transport and congestion-control protocols, and other network systems and applications.
I completed my Ph.D. in computer science at Stanford University, advised by
Keith Winstein and
Philip Levis.
My dissertation was
on the development of platforms and algorithms to achieve
practical reinforcement learning (RL) on the
Internet, in the context of video streaming and congestion control.
My work has received the
Applied Networking Research Prize,
the
USENIX NSDI Community Award, and the
USENIX ATC Best Paper Award. I am the creator of
Puffer, a live-streaming site
that has reached 150,000 real users (as of June 2021).
Before that, I graduated from Tsinghua University, where I received
a B.S. in computer science from
Yao Class (founded by Turing Award laureate
Andrew Yao)
and a B.A. in economics. I also studied at MIT in 2014.
Here is my CV.
I tweet research internship opportunities and bad jokes on Twitter.
Sept
2021
We have a research internship opening for an RL-in-videoconferencing project I have
been leading. Due to some special arrangement of this collaborative project between
MSR Redmond & Asia, applicants must physically be present in China and will
work from MSR Asia.
May
2021
My colleague Mike Liang at MSR Asia and I are seeking a research intern to work
with us on microservices. The internship is affiliated with MSR Asia and requires
the intern to be physically located in Beijing. Female candidates are encouraged
to apply! Update: We recruited a female intern!
Mar
2021
Check out the simple
Chrome extension
I developed to decode and display
Microsoft Safe Links
in a web page or email as the original URLs, making the links easier on the eyes
while preserving the enhanced security.
Zhengxu Xia*, Yajie Zhou*, Francis Y. Yan, Junchen Jiang (*equal contribution)
In submission; draft available upon request.
arrow_right
We proposed Genet, a novel training framework that enhances the generalization of
reinforcement learning (RL) policies in systems and networking. In combining the
latest RL generalization technique with judicious use of rule-based heuristics,
Genet substantially improves the performance of simulation-trained RL policies
under unseen workloads and in real environments.
Michael Rudow, Francis Y. Yan, Ganesh Ananthanarayanan,
Martin Ellis, K.V. Rashmi
In submission; draft available upon request.
arrow_right
We designed Tambur, a new approach to forward error correction (FEC) for
videoconferencing built upon streaming codes and machine learning.
Compared with the FEC scheme deployed in production by a commercial
videoconferencing application, Tambur reduced decoding failures of video frames
by 26% while employing 35% less bandwidth for redundancy.
Francis Y. Yan, Hudson Ayers, Chenzhi Zhu, Sadjad Fouladi, James Hong,
Keyi Zhang, Philip Levis, Keith Winstein
USENIX Symposium on Networked Systems Design and Implementation (NSDI),
February 2020
arrow_right
We built
Puffer, a free, publicly
accessible website that live-streams television channels and operates as a
randomized experiment of adaptive bitrate (ABR) algorithms.
As of June 2020, Puffer has attracted 120,000 real users and streamed 60 years
of video across the Internet. Using Puffer, we developed an ML-based ABR algorithm,
Fugu, that robustly outperformed existing schemes by learning
in situ,
on real data from its actual deployment environment.
Francis Y. Yan, Jestin Ma, Greg D. Hill, Deepti Raghavan, Riad S. Wahby,
Philip Levis, Keith Winstein
USENIX Annual Technical Conference (ATC), July 2018
arrow_right
Pantheon is a “training
ground” for congestion-control research and has
assisted four schemes from other research groups in publishing at
NSDI 2018
(
Copa and
Vivace),
ICML 2019
(
Aurora),
and SIGCOMM 2020 (
TCP-TACK).
It also enabled our own ML-based
congestion-control algorithm, Indigo, which was trained to imitate expert
congestion-control algorithms we created in emulation
and achieved good performance over the real Internet.
- Applied Networking Research Prize
(for Puffer), Internet Research Task Force (IRTF), 2021
- USENIX NSDI Community Award (for Puffer),
2020
- USENIX ATC Best Paper Award (for Pantheon),
2018
- Award of Excellence, Stars of Tomorrow Internship Program,
Microsoft Research, 2015
- Outstanding Graduate of Tsinghua University
(lone recipient at my institution), 2015
- Outstanding Graduate of Beijing, China, 2015
- Merrill Lynch Fellowship, Massachusetts Institute of Technology
(declined), 2015
- National Scholarship, China, 2014
- Silver Prize, Yao Award, Tsinghua University, 2014
- Tsinghua University Comprehensive Scholarship, 2013
- National Endeavor Fellowship, China, 2012
- Tsinghua-Baidu Scholarship, 2012
- Scholarship for Tsinghua Xuetang Talents Program, 2011–2014
- First Prize, National Senior High School Mathematical Olympiad, China, 2010
- First Prize, National Olympiad in Informatics in Provinces, China, 2008
-
External Reviewer, NSDI 2022
-
Editorial Board Member, Journal of Systems Research (2021-2022)
-
Challenge Chair, ACM MMSys 2021
-
Organizer,
Grand Challenge on Banwidth Estimation for Real-Time Communications,
ACM MMSys 2021
-
Reviewer, IEEE/ACM Transactions on Networking (2019)
-
Reviewer, ACM SIGCOMM Computer Communication Review (2019)
-
Reviewer, Computer Communications (2019)
