
Linguists estimate that at least half of the world's estimated 7,000 spoken languages will become extinct by the century's end, due to forces ranging from globalization to cultural assimilation.
Part of the challenge of documenting and revitalizing endangered languages is a lack of texts and speech recordings to work with. Seneca, a language of one of the six Iroquois Nations in North America, has only about 100 first-language speakers and several hundred more second-language learners.
Automatic speech recognition (ASR) technology is widely used to transcribe languages with millions or billions of speakers, like English and Mandarin. But it has only scratched the surface with languages like Seneca, which have vastly fewer speakers and significantly less data to work with.
Now a team of researchers at the Rochester Institute of Technology in New York, along with colleagues from the University at Buffalo, is tapping deep learning to bolster the ability of ASR. And while its focus is on Seneca, the project's vision encompasses the preservation of languages globally as well as an important part of our shared cultural history.
Knowing about different languages teaches us a lot about how our brain works, said Emily Prud'hommeaux, an assistant professor of computer science at Boston College and a research faculty member at RIT. When you document a language, you're preserving information not only about that language but also about how humans use language in general.
It's no coincidence that Prud'hommeaux and her team started with the Seneca language. Three members of the Seneca nation are part of the effort - a direct connection that is rare in research of this type, she said.
Leading the charge is Robbie Jimerson, a Ph.D. student in RIT's Golisano College of Computing and Information Science. He is a member of the Seneca Nation of Indians and is passionate about ensuring the survival of the Seneca language.
There's a big effort by the leaders of the tribe to preserve and promote our language, said Jimerson. I was looking for an opportunity to contribute.
Using GANs to Create More Language Samples Now in its third year, the project has had challenges when it comes to accumulating language data. Jimerson said the Seneca community can be guarded about what it shares with other people, so there wasn't an abundance of recordings of the language being spoken. He set out to change that.
He started by recording friends and elders who speak the language and asking them to record their friends. He found out whenever someone was speaking Seneca in public. He asked for family recordings of elders telling stories handed down from previous generations. And he grabbed any publicly available videos or recordings he could find online.
The team has fine-tuned an ASR model for Seneca, running it through generative adversarial networks to create more samples out of the limited number of recordings. The model turns wave files of the spoken language into streams of characters, while computing probability and making corrections.
The resulting data is fed into a deep learning model that in turn expands upon the ASR model's accuracy.
The team's networks run in two compute settings: on a nine-server machine learning lab running a variety of NVIDIA Tesla GPUs, and on a university cluster of large servers, each running 10 NVIDIA Tesla P4 GPUs. Each cluster runs a range of deep learning frameworks such as TensorFlow and Caffe.
The computer engineering cluster is for all students in the computer engineering department, and so they have to compete' for these resources, said Ray Ptucha, assistant professor of computer engineering at RIT, another collaborator on this project.
With access to these clusters at a premium, Jimerson tests code and checks the stability of models on a local machine running an NVIDIA TITAN X rather than inconvenience other students by running a model that might crash.
Achieving Better Accuracy So far, the team's efforts have brought the word error rate of its ASR model from 70 percent down to 56 percent. The goal, said Prud'hommeaux, is to get that rate down to 25 percent, which is where ASR systems were in processing English several years ago.
The more samples of spoken and written Seneca the team can accumulate, the more the error rate will decrease. (Today, English ASR models can achieve word error rates as low as 5 percent.)
The team's work is expected to help with language preservation efforts around the world.
Prud'hommeaux said the team has an agreement with an archiving institution that's a condition of a grant the project received from the National Science Foundation. The resulting language archiving database will be made available as a resource for other efforts seeking to document threatened languages.
Additionally, Prud'hommeaux said the team's work could prove helpful for any deep learning effort that has to make do with limited amounts of data.
Read more about the team's work in their research papers here and here.
Feature image: The Haudenosaunee (Iroquois Confederacy) flag, via Wikimedia Commons.
Most recent headlines
06/10/2025
France T l visions, France's leading broadcaster, has received the 2025 EBU ...
04/09/2025
Monumental Sports & Entertainment (MSE), in collaboration with Dalet, has been a...
04/07/2025
By Lucy Spicer
Editor's note: We are deeply saddened by the passing of Robe...
04/07/2025
The Summer Reading Challenge has inspired children across the U.K. to embrace a ...
04/07/2025
Introducing CVP Auctions: Trusted Tech, At Auction-Only Prices
Brie Clayton July 4, 2025
0 Comments
A New Destination for Film and Broadcast Gear at U...
04/07/2025
Unveiling China's Ecological Wonders with URSA Mini Pro 12K
Brie Clayton July 4, 2025
0 Comments
Nature Documentaries Freeze Epic Momentsin Extrem...
04/07/2025
Create Any Planet in After Effects
Graham Quince July 4, 2025
0 Comments
The Advanced 3D renderer in After Effects makes it really easy to import 3D o...
04/07/2025
Custom Gradient Colorize for Resolve Fusion
Simon Ubsdell July 4, 2025
0 Comments
Digging into the theory of the gradient colorize process to build ou...
04/07/2025
Watching live sports on TV these days is a delicate balance for broadcasters how much information is too much?...
04/07/2025
WASHINGTON As Federal Communications Commission Chair Brendan Carr works to achieve his goal of a very busy, productive summer at the FCC, the agency has rele...
04/07/2025
Other World Computing Launches OWC Guardian USB-C Hardware-Encrypted Portable SS...
04/07/2025
CANAL Brings MotoGP to Vision Pro with Blackmagic URSA Cine Immersive
Brie Clayton July 3, 2025
0 Comments
Experience MotoGP like never before as CA...
04/07/2025
Friday 4 July 2025
Watch the trailer here https://youtu.be/Od5WpvWmDr8
Further...
04/07/2025
The Pitch comes to RT Radio 1 airwaves this Sunday at 7pm. The Pitch is a brand...
03/07/2025
For our Vario-Jib, as well as for a repertoire of weights, we now have carts that make transport easier...
03/07/2025
One of our latest innovations: Slider Drive Unit!
Our Vario Slider not only comes with improvements and new features, but also with its own drive unit!
Our i...
03/07/2025
The 360 EVO lives up to its name: 360-degree rotation not only on the pan axis, but also on the roll axis, enables filmmakers to take their creative visions eve...
03/07/2025
What a buzzing few days at BSCExpo in London!
We were thrilled to connect with so many talented professionals from the world of cinematography.
A huge thank ...
03/07/2025
ADOLESCENCE - Shot with Panther Tower"The whole rig had to be detachable mid-episode
Netflix limited drama series Adolescence has been gaining attentio...
03/07/2025
To achieve a true level of decisive overmatch, communication systems must be as ...
03/07/2025
The Broadcasters Foundation of America has named Ray Hopkins, president of U.S. distribution at Paramount Global, to its board of directors....
03/07/2025
SIOUX FALLS, S.D. In his first major policy speech, Federal Communications Commission chair Brendan Carr outlined his Build America Agenda for the agency'...
03/07/2025
As the industry is making a major push to transition to NextGen TV and develop new businesses that will fund the next generation of broadcasting, TV Tech's ...
03/07/2025
Despite enabling its subscribers to watch films and TV shows in the highest possible resolution, Netflix has revealed it is taking things old school with the in...
03/07/2025
WASHINGTON In response to news that Paramount Global has agreed to pay $16 million to settle a lawsuit by President Donald Trump, Democrats and other administra...
03/07/2025
WUPPERTAL, Germany Riedel Communications today launched RefSuite for professional sports workflows....
03/07/2025
WEITERSTADT, Germany LYNX Technik, a provider of modular signal processing solutions, has announced that founder and longtime CEO Winfried Deckelmann is steppin...
03/07/2025
WASHINGTON Representatives from six trade groups representing tech, pay TV, broadband, LPTV and other sectors recently met with Federal Communications Commissio...
03/07/2025
PLYMOUTH, Wis. Dan Whealy has acquired Heartland Video Systems and will be its new owner and president, the systems integrator said....
03/07/2025
Acclaimed filmmaker, educator, and industry veteran Jay Holben has released his most ambitious work to date: The Director's Guide to Everything a sweeping...
03/07/2025
For Los Angeles-based photographer Scout Pruski, documenting the annual Dyke Day LA is a way to honor and give back to the city's legacy queer community and...
03/07/2025
Modular, Travel-Ready Solution Matthews Studio Equipment launches the new Chimera V-Flat Panel Frame Kit, the lightweight and modular, ultra-durable alternativ...
03/07/2025
Dalet, a leading technology and service provider for media-rich organizations, today announced the release of Dalet Flex 2025.6 LTS (long-term support), introdu...
03/07/2025
Lightware's LARA Wizard makes it quicker and easier than ever to set up audio-visual (AV) room automation and control configurations, ensuring anyone, from ...
03/07/2025
Alfalite, the only European manufacturer of LED screens, has obtained both the NCAGE Code (NATO Commercial and Government Entity) and TAA (Trade Agreements Act)...
03/07/2025
nxtedition is implementing a new solution for WIN Network, based in its headquarters in Wollongong, New South Wales, Australia. The nxtedition integrated platfo...
03/07/2025
ORS Group and Insys Video Technologies today announce the launch of a new joint brand identity: Big Blue Marble. The move reflects the evolution of a long-stand...
03/07/2025
Evergent, a global leader in customer lifecycle management and monetization solutions for subscription-based businesses, today announced the successful results ...
03/07/2025
LYNX Technik, provider of modular signal processing solutions today announces an executive leadership transition, effective July 1, 2025. Markus Motzko has been...
03/07/2025
Chaos Releases V-Ray for Blender
Brie Clayton July 2, 2025
0 Comments
Production-Proven Renderer V-Ray Arrives Natively in Blender for the First Time;...
03/07/2025
Calvin University Adds Blackmagic Replay to Broadcasts and Streams
Brie Clayton July 2, 2025
0 Comments
Replay workflow joins 14 Blackmagic Design cam...
03/07/2025
Berklee's Gospel Performance Program Welcomes Jonathan Nelson as the Herb Al...
03/07/2025
Back to All News
Get Ready for Even More Action! Netflix Has Confirmed the Thir...
03/07/2025
Back to All News
Get ready for even more action! Netflix has confirmed the thir...
03/07/2025
In today's fragmented media landscape, advertisers expect more than just reach - they want seamless cross-platform campaigns that deliver measurable results...
03/07/2025
** MEDIA ALERT ** Shonen Jump Store Hosting Los Angeles Pop-up During Anime Exp...
03/07/2025
Thursday 3 July 2025
David Harewood and Matthew Broome in Mozart: Genius for Hi...
03/07/2025
The Sky WNT Fund supports Womens National Team players with a 30,000 bursaryThursday 3 July 2025
Sky Ireland has announced the four recipients of the 2025 Sky...
03/07/2025
The forecast this month is showing a 100% chance of epic gaming. Catch the scorching lineup of 20 titles coming to the cloud, which gamers can play whether indo...
02/07/2025
Hege Wik and Odin appear in FOLKTALES by Heidi Ewing and Rachel Grady, an official selection of the 2025 Sundance Film Festival. Courtesy of Sundance Institut...