Sony Pixel Power calrec Sony

How Deep Learning Is Aiding Preservation of Seneca and Other Endangered Languages

03/01/2019

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.
LINK: https://blogs.nvidia.com/blog/2019/01/02/deep-learning-preserves-senec...
See more stories from nvidia

Most recent headlines

22/11/2025

IDX Unveils CUE-J Series Batteries

CARSON, Calif. IDX has introduced the IDX CUE-J Series battery/charger kits, including the CUE-J98, CUE-J150 and CUE-J198....

22/11/2025

NBA: More Than 60 Million Watched Games in First Month, Best in 15 Years

The NBA has released encouraging viewing and social media data that the beginnings of its $76 billion deal with NBC/Peacock, Prime Video and ESPN are paying off...

22/11/2025

FCC Sets Deadlines for Comments on New NextGen TV Proposals

WASHINGTON The Federal Communications Commission has set deadlines for comments on its newest proposals for NextGen TV, aka ATSC 3.0, with comments due on Jan. ...

22/11/2025

Seeking Advice for a New Opera, Laura Kaminsky Consulted the Experts: Her Students

Seeking Advice for a New Opera, Laura Kaminsky Consulted the Experts: Her Studen...

21/11/2025

Platinum White Paper: Appear Shares Why Media Exchange Is the Missing Link in Software-Defined Live Production

Platinum White Paper: Appear Shares Why Media Exchange Is the Missing Link in So...

21/11/2025

NWSL Championship 2025: CBS Sports To Deploy Two-Point FlyCam for Match Coverage at PayPal Park

NWSL Championship 2025: CBS Sports To Deploy Two-Point FlyCam for Match Coverage...

21/11/2025

NWSL Caps 2025 Season With Awards Show, Skills Challenge Productions

NWSL Caps 2025 Season With Awards Show, Skills Challenge ProductionsA team of 70 is on the ground in California to produce both eventsBy Mark J Burns, SVG Contr...

21/11/2025

USL and NEP Ready for Largest USL Championship Final Production Ever

USL and NEP Ready for Largest USL Championship Final Production EverThe broadcast from Tulsa, OK, will air CBS and TUDN on Saturday at 12 p.m. ETBy Jason Dachma...

21/11/2025

With Two New Teams, PWHL Boosts Production Workforce and Central Review for Season 3

With Two New Teams, PWHL Boosts Production Workforce and Central Review for Seas...

21/11/2025

Dinner and a Movie: Jared Lank on Powwow Highway and Luskinikn

Jared Lank and his mother in the '90s...

21/11/2025

Explore the Lands of Oz on Spotify With This Exclusive Wicked' Experience

Fans have been counting down the days until the final theatrical chapter of Wicked is revealed. To celebrate the highly anticipated release of Wicked: For Good ...

21/11/2025

Spotify House Seoul Delivered Unforgettable Performances From Central Cee, The Kid LAROI, Jay Park, ZICO, GroovyRoom, and More

Last week, Spotify turned up the volume in Seoul with the return of Spotify Hous...

21/11/2025

Training success at SGL Carbon in Meitingen

Wiesbaden, November 21, 2025. The SGL Carbon site in Meitingen has reason to celebrate as one of its trainees received a special award. Elias Stemmer was honore...

21/11/2025

L3Harris Recognizes Employee Achievements with LHX Excellence Awards

MELBOURNE, Fla., Nov. 21, 2025 - L3Harris Technologies (NYSE: LHX) has announced this year's LHX Excellence Awards, the company's most prestigious recog...

21/11/2025

FCC Proposes Upper C-Band Rules for 2027 Auction

WASHINGTON The Federal Communications Commission by a 3-0 vote opened a notice of proposed rulemaking (NPRM) to advance Congress's mandate to clear a minimu...

21/11/2025

FCC Votes to Clear at Least 100MHz of Upper C-Band Spectrum

WASHINGTON The Federal Communications Commission by a 3-0 vote adopted a Notice of Proposed Rulemaking (NPRM) to advance Congress's mandate to clear a minim...

21/11/2025

Spectrum Expands 4K Content to Apple TV 4K and Roku Devices

STAMFORD, Conn. Charter's Spectrum has expanded the devices that can offer 4K content on the Spectrum TV app to compatible Apple TV 4K and Roku devices....

21/11/2025

Study: Salaries in Content, Connectivity Industries Continue To Grow

NAPERVILLE, Ill. Media industry employers are continuing their multiyear trend of increasing salaries for all worker segments but lag general industry raises, s...

21/11/2025

NAB Opens Nominations for 2026 Technology Awards

WASHINGTON The National Association of Broadcasters said it is accepting nominations for the 2026 NAB Technology Awards, honors that recognize excellence in bro...

21/11/2025

AAT Introduces Automated RF Line Analysis

American Amplifier Technologies has released a vector network analysis module....

21/11/2025

The Best Movie Musicals on Every Streaming Platform

The Best Movie Musicals on Every Streaming Platform From Wicked to The Sound of Music, heres where to stream all the classic movie musicals and recent hits on...

21/11/2025

Space42 and Hisdesat Lay Groundwork for UAE-Spain Partnership to Advance Satellite Innovation and Secure Communications

The agreement creates a platform for joint collaboration, technology integration...

21/11/2025

GoodGym named Grand Prix winner of the 2025 Sky Zero Footprint Fund

Sky Media's £2m award-winning sustainability initiative crowns its first charity as this year's standout changemakerFriday 21 November 2025 GoodGym nam...

21/11/2025

Standards Pavilion concludes COP30 with call to embed International Standards in every step of climate action

As COP30 draws to a close, the International Electrotechnical Commission (IEC), ...

20/11/2025

MLB Media-Rights Shakeup: NBC's New Three-Year Deal Covers Sunday Night Baseball,' Peacock-Exclusive Games, and More

MLB Media-Rights Shakeup: NBC's New Three-Year Deal Covers Sunday Night Bas...

20/11/2025

MLB Media-Rights Shakeup: New Deal Will Bring 30 National Games to ESPN's Linear Platform, MLB.TV Exclusively to ESPN App

MLB Media-Rights Shakeup: New Deal Will Bring 30 National Games to ESPN's Li...

20/11/2025

MLB Media-Rights Shakeup: Netflix Lands Opening Night, Home Run Derby, Field of Dreams

MLB Media-Rights Shakeup: Netflix Lands Opening Night, Home Run Derby, Field of ...

20/11/2025

MLB Media-Rights Shakeup Overview: ESPN, NBCU, Netflix Ink Three-Year Deals

MLB Media-Rights Shakeup Overview: ESPN, NBCU, Netflix Ink Three-Year DealsESPN gets new 30-game package, MLB.TV; NBC extends Sunday nights; Netflix adds tentpo...

20/11/2025

SVG Students To Watch: Henry Thuss, Indiana University

SVG Students To Watch: Henry Thuss, Indiana UniversityThe Southern California product has his goals set on the front benchBy Brandon Costa, Director of Digital ...

20/11/2025

Done+Dusted's Guy Carrington on Creating the Spectacular League of Legends World Championship Opening Ceremony

Done+Dusted's Guy Carrington on Creating the Spectacular League of Legends W...

20/11/2025

FIA Extreme H World Cup Host Broadcaster Aurora Goes Inside the Production of the Hydrogen-Fuelled Motorsport

FIA Extreme H World Cup Host Broadcaster Aurora Goes Inside the Production of th...

20/11/2025

Platinum White Paper: Amagi Utilizes Cloud Production for Sports Events - Multi-Camera Live Workflow with Remote Commentary and Graphics

Platinum White Paper: Amagi Utilizes Cloud Production for Sports Events - Multi-...

20/11/2025

2025 Sports Broadcasting Hall of Fame: Marc Herklotz, Steady Hand Behind the Scenes

2025 Sports Broadcasting Hall of Fame: Marc Herklotz, Steady Hand Behind the Sce...

20/11/2025

NFL Deep Dive: How 32 Cameras at Each Stadium Drive Virtual Measurement, Boundary Replays, and Skeletal Tracking

NFL Deep Dive: How 32 Cameras at Each Stadium Drive Virtual Measurement, Boundar...

20/11/2025

Zodiac Killer Project Refuses to Kill Its Darlings With an Avant-Garde Essay Film

Charlie Shackleton attends the 2025 Sundance Film Festival premiere of Zodiac K...

20/11/2025

5 Ways to Get More out of Your Playlists on Spotify

Your playlists are personal. They're the soundtracks to your road trips, your quiet mornings, and your biggest celebrations; collections of memories and dis...

20/11/2025

Spotify, Trk mziinin efsanelerine sayg duruu niteliindeki ICON Trkiye'yi duyurdu

Spotify, uzun s redir zerine al t T rk m zik k lt r n n ikon haline gelmi ...

20/11/2025

Spotify and The Hollywood Reporter' Partner on the First-Ever Podcaster Roundtable

For the first time, Spotify has teamed up with The Hollywood Reporter to cohost ...

20/11/2025

Spotify and the ARIA Awards Mark a New Era for Australian Music

This year's ARIA Awards marked a turning point for Australian music, and Spotify was right at the heart of it. For the first time in the awards' nearly ...

20/11/2025

Cracking the Code for Resilient, Secure Space-based Communications

L3Harris has achieved NSA Cybersecurity Directorate certification for its KSV-650 space hub end cryptographic unit, ensuring secure, adaptable communications fo...

20/11/2025

Continued Commitment and Investment in Poland

Left to Right: David Taubman, Regional Managing Director, Central and Eastern Europe; Arek Szalpuk, Poland Sr. Account Manager; Mr. Marcin Wi niewski, President...

20/11/2025

Cinegy wins the Swedish Hockey League

15 March 2012 SHARE Facebook Twitter Linkedin Email Cinegy and Vericom are pleased to announce the recent deployment of Cinegy Archive for production and as...

20/11/2025

2012

6 September 2012 SHARE Facebook Twitter Linkedin Email FIC Turkey supplies media content to Pay TV under Fox TV Channel brands in Turkey. At the start of s...

20/11/2025

VIVA Indonesia broadcasting the World Cup Brazil 2014 with Cinegy

25 June 2014 SHARE Facebook Twitter Linkedin Email Last month VIVA finalized its Cinegy installation in order to produce and broadcast the World Cup Brazil ...

20/11/2025

2014

1 October 2014 SHARE Facebook Twitter Linkedin Email With 5 locations spread out around Baden-W rttemberg, Schw bisch Media required a scalable end to end s...

20/11/2025

Strictly Come Dancing Goes with a Cinegy Workflow

27 April 2015 SHARE Facebook Twitter Linkedin Email Cinegy, which develops and produces media asset management products that are used in flagship production...

20/11/2025

Cinegy announces new Latin America partnership

12 August 2015 SHARE Facebook Twitter Linkedin Email Munich, Germany, 26 October 2015 - Cinegy, which develops and produces software technology for digital ...

20/11/2025

2015

5 November 2015 SHARE Facebook Twitter Linkedin Email Munich, Germany 3 November 2015 - Cinegy, which develops and produces software technology for digital ...

20/11/2025

Cinegy launches Cinegy Open' initiative

12 February 2016 SHARE Facebook Twitter Linkedin Email Munich, Germany 9 February 2016 - Cinegy, which develops and produces software technology for digital...