How MIT’s Muriel Medard Pioneered the Universal Decoder

It is expected to increase the efficiency of AR devices and 5G

4 min read
Photo of a woman behind a blue background.
MURIEL MEDARD

Network coding pioneer Muriel Médard calls herself a pathological optimist. The MIT electrical engineering and computer science professor’s positive thinking has led to new ways to improve tried-and-true techniques in the field of information theory.

As the head of the network coding group at the university’s Research Laboratory for Electronics, the IEEE Fellow led a team that created a silicon chip that eliminates the need for custom decoding hardware to spot signal errors. The chip uses a new algorithm the team developed with Ken Duffy from Maynooth University: guessing random additive noise decoding (GRAND). Duffy is a professor and the director of the school's Hamilton Institute.


“A lot of other coding mechanisms are really geared toward the worst-case [scenario]—which is probably too pessimistic but also not energy efficient and not time efficient,” Médard says.

Médard is this year’s recipient of the IEEE Koji Kobayashi Computers and Communications Award “for contributions to the theory and practice of network coding, optical networks, and wireless communications.” NEC sponsors the award.

She says receiving the award, which recognizes those who integrate computers and communications, is “pretty amazing” because her work tends to be more on the theoretical side. She likes to combine different fields, she says, because “there’s a richness and a set of really cool things you can do in between.”

“Once you’ve seen certain types of techniques and certain ways of thinking [from other fields], you can’t unlearn them,” she says. “The award really felt like a validation of this approach that I have taken—not in a calculated, career-advancing way but in an injudicious, overly optimistic way.”

MANY OPTIONS

Médard says she didn’t know what she wanted to be growing up. She liked working on creative projects and mathematics, so she decided to study math and literature at MIT. She had not been exposed to engineering before MIT, but she took a class and discovered her passion.

“There were many different fields where I could see myself being happy, but [engineering] is just such an interesting field,” she says. “It’s dynamic, and it offers so many creative possibilities. It’s never boring.”

Médard earned all five of her degrees at MIT: bachelor’s degrees in electrical engineering and computer science, mathematics, and humanities, plus master’s and doctoral degrees in EE.

After graduating, she joined the University of Illinois at Urbana–Champaign in 1998 as an assistant professor. She returned to MIT as a faculty member in 2000.

Aspects of networking, network communications, and network computing—especially coding systems—have been at the core of her work.

She holds more than 50 U.S. and international patents, most of which have been licensed or acquired.

She helped to foundtwo companies to commercialize some of the coding systems she helped invent.

She is a consultant for Code On, an MIT spin-out in Cambridge, Mass., that was founded by Médard and the other inventors of random linear network coding (RLNC). Among other things, RLNC can help speed up the Internet and improve video quality for streaming movies and live events.

She is chief scientist at Steinwurf, the other company she helped found. Based in Aalborg, Denmark, Steinwurf provides correction products using RLNC. The technology creates repair packets, which can be used to recreate lost data.

ELIMINATING NOISE

Médard’s GRAND chip is the result of a collaborative effort by MIT, Boston University, and Maynooth University. In addition to decoding just about any error-correcting codes, it can decode ones that don’t yet exist, Médard says.

Error correction can protect data sent over the Internet—including email, text messages, and images—from being damaged by electromagnetic interference.

To address data protection, Médard says, the usual approach has been to add a code consisting of a pattern of 1s and 0s to the sequence of transmitted signals. Those patterns are then used at the receiver to examine the signal for errors and try to reconstruct what was transmitted.

Each code is designed to correspond with a decoding algorithm that can be computationally complex. The decoders use a codebook to try to determine what the transmitted information might be. Each codebook requires a separate chip or other piece of hardware.

The GRAND chip, which was designed in 40-nanometer CMOS technology, eliminates the need for code-specific decoders and enables universal decoding that rapidly cycles through all plausible noise patterns. The chip can process a high volume of data with little lag time. The universal decoder is expected to increase the efficiency of devices that process high volumes of data such as augmented and virtual-reality devices, gaming systems, and 5G networks.

Médard says solving the decoding problem was a great example of taking a problem that has been considered since the inception of the field and thinking about it in a different, more optimistic way.

“The approach we took was looking at the receiver side,” she says. “GRAND decodes all the codes you might want to put in there, including codes that up until now had no decoder.”

HOPEFUL FOR THE FUTURE

Médard is an active IEEE volunteer. She served as the 2012 president of the IEEE Information Theory Society. While a member of the society’s board of governors, she established a popular mentoring program to connect senior and junior researchers.

A former editor in chief of the IEEE Journal on Selected Areas in Communications, last year she became editor in chief of the IEEE Transactions on Information Theory.

She has presented at numerous IEEE conferences, and many of her research articles are available in the IEEE Xplore Digital Library.

She volunteers, she says, because she believes a professional society is important to the engineering field.

“Having conferences at which we exchange information is key,” she says. “That’s where you present to your colleagues, you get feedback, and you solidify and ascertain the validity of your work. The same with the journals: They are the archival repository of our collective progress.”

IEEE has a role to play beyond exchanging information, she says.

“This is how we are showing the world what we think is worthwhile and is valuable about our profession,” she says. “This is part of not just getting recognition from colleagues but also part of our messaging to the wider world. There’s so much discussion about the role and sometimes even the value of technology. It’s crucial that as a profession we are able to articulate that value in a compelling way.”

The Conversation (0)

Get unlimited IEEE Spectrum access

Become an IEEE member and get exclusive access to more stories and resources, including our vast article archive and full PDF downloads
Get access to unlimited IEEE Spectrum content
Network with other technology professionals
Establish a professional profile
Create a group to share and collaborate on projects
Discover IEEE events and activities
Join and participate in discussions

Deep Learning Could Bring the Concert Experience Home

The century-old quest for truly realistic sound production is finally paying off

12 min read
Vertical
Image containing multiple aspects such as instruments and left and right open hands.
Stuart Bradford
Blue

Now that recorded sound has become ubiquitous, we hardly think about it. From our smartphones, smart speakers, TVs, radios, disc players, and car sound systems, it’s an enduring and enjoyable presence in our lives. In 2017, a survey by the polling firm Nielsen suggested that some 90 percent of the U.S. population listens to music regularly and that, on average, they do so 32 hours per week.

Behind this free-flowing pleasure are enormous industries applying technology to the long-standing goal of reproducing sound with the greatest possible realism. From Edison’s phonograph and the horn speakers of the 1880s, successive generations of engineers in pursuit of this ideal invented and exploited countless technologies: triode vacuum tubes, dynamic loudspeakers, magnetic phonograph cartridges, solid-state amplifier circuits in scores of different topologies, electrostatic speakers, optical discs, stereo, and surround sound. And over the past five decades, digital technologies, like audio compression and streaming, have transformed the music industry.

Keep Reading ↓Show less
{"imageShortcodeIds":[]}