Advancing AI requires a full-stack approach, with a powerful foundation of computing infrastructure - including accelerated processors and networking technologies - connected to optimized compilers, algorithms and applications.
NVIDIA Research is innovating across this spectrum, supporting virtually every industry in the process. At this week's International Conference on Learning Representations (ICLR), taking place April 24-28 in Singapore, more than 70 NVIDIA-authored papers introduce AI developments with applications in autonomous vehicles, healthcare, multimodal content creation, robotics and more.
ICLR is one of the world's most impactful AI conferences, where researchers introduce important technical innovations that move every industry forward, said Bryan Catanzaro, vice president of applied deep learning research at NVIDIA. The research we're contributing this year aims to accelerate every level of the computing stack to amplify the impact and utility of AI across industries.
Research That Tackles Real-World Challenges Several NVIDIA-authored papers at ICLR cover groundbreaking work in multimodal generative AI and novel methods for AI training and synthetic data generation, including:
Fugatto: The world's most flexible audio generative AI model, Fugatto generates or transforms any mix of music, voices and sounds described with prompts using any combination of text and audio files. Other NVIDIA models at ICLR improve audio large language models (LLMs) to better understand speech.
HAMSTER: This paper demonstrates that a hierarchical design for vision-language-action models can improve their ability to transfer knowledge from off-domain fine-tuning data - inexpensive data that doesn't need to be collected on actual robot hardware - to improve a robot's skills in testing scenarios.
Hymba: This family of small language models uses a hybrid model architecture to create LLMs that blend the benefits of transformer models and state space models, enabling high-resolution recall, efficient context summarization and common-sense reasoning tasks. With its hybrid approach, Hymba improves throughput by 3x and reduces cache by almost 4x without sacrificing performance.
LongVILA: This training pipeline enables efficient visual language model training and inference for long video understanding. Training AI models on long videos is compute and memory-intensive - so this paper introduces a system that efficiently parallelizes long video training and inference, with training scalability up to 2 million tokens on 256 GPUs. LongVILA achieves state-of-the-art performance across nine popular video benchmarks.
LLaMaFlex: This paper introduces a new zero-shot generation technique to create a family of compressed LLMs based on one large model. The researchers found that LLaMaFlex can generate compressed models that are as accurate or better than state-of-the art pruned, flexible and trained-from-scratch models - a capability that could be applied to significantly reduce the cost of training model families compared to techniques like pruning and knowledge distillation.
Proteina: This model can generate diverse and designable protein backbones, the framework that holds a protein together. It uses a transformer model architecture with up to 5x as many parameters as previous models.
SRSA: This framework addresses the challenge of teaching robots new tasks using a preexisting skill library - so instead of learning from scratch, a robot can apply and adapt its existing skills to the new task. By developing a framework to predict which preexisting skill would be most relevant to a new task, the researchers were able to improve zero-shot success rates on unseen tasks by 19%.
STORM: This model can reconstruct dynamic outdoor scenes - like cars driving or trees swaying in the wind - with a precise 3D representation inferred from just a few snapshots. The model, which can reconstruct large-scale outdoor scenes in 200 milliseconds, has potential applications in autonomous vehicle development.
Discover the latest work from NVIDIA Research, a global team of around 400 experts in fields including computer architecture, generative AI, graphics, self-driving cars and robotics.
North America Stories
25/04/2025
YouTube is celebrating its 20th birthday with stats showing how popular the platform has become and by announcing some new features, including the ability for u...
25/04/2025
BALTIMORE Sinclair, Inc. has announced the retirement of Dave Schwartz, its corporate senior vice president/station operations, effective June 30....
25/04/2025
The Society of Broadcast Engineers (SBE) has released its popular CertPreview practice exams in a new, online format. SBE CertPreview allows people preparing to...
25/04/2025
ARLINGTON, Texas Two months after naming Tegna station KFAA in Dallas-Fort Worth its exclusive local broadcast partner for WNBA games, the Dallas Wings are expa...
25/04/2025
LOS ANGELES, Calif. Allen Media Group has announced that all 223 U.S. Markets covered by its Local Now FAST Channel are now available on the Vizio WatchFree+ st...
25/04/2025
Working with Shutterstock's AI Editor to Create Unique Images
Brie Clayton April 25, 2025
0 Comments
What's better than a stock image source? ...
24/04/2025
We're huge fans of shining a spotlight on the documentaries that empower, en...
24/04/2025
Highlights*
Revenue of $5.1 billion
Operating margin of 10.2%; Adjusted segment operating margin of 15.6%
Diluted EPS of $2.04; Non-GAAP diluted EPS of $2.4...
24/04/2025
For more than 50 years, the company's commercial approach to delivering resi...
24/04/2025
Sonnet Announces Solo5G USB-C to 5 Gigabit Ethernet Adapter
Brie Clayton April 24, 2025
0 Comments
Compact, Bus-powered Adapter Adds Instant 5 Gigabit...
24/04/2025
Sony Electronics Launches FE 50-150MM F2 GM
Brie Clayton April 24, 2025
0 Comments
The World's Firsti Telephoto Zoom Lens with a Maximum Focal Len...
24/04/2025
Branded cinematic shorts and documentaries have surged in popularity as more companies strategize out-of-the-box ways to connect and engage with audiences. Whil...
24/04/2025
Compact, Bus-powered Adapter Adds Instant 5 Gigabit Ethernet Connectivity to Computers With USB-C or Thunderbolt ports
What's New:
Sonnet Technologies tod...
24/04/2025
Black Box , a leading digital infrastructure solution provider, today announced that its Emerald DESKVUE PE is a remote production category winner in the 2025 ...
24/04/2025
Interra Systems, a leading provider of end-to-end quality assurance solutions for the digital media industry, today announced that its ORION content monitoring ...
24/04/2025
FingerWorks Telestrators continues to be at the forefront of live sports broadcasting with its cutting-edge solutions, supporting major events like NASCAR, the...
24/04/2025
IBC announces the launch of the IBC2025 Innovation Awards, which recognise pioneering advances in technology and social impact in the media and entertainment (M...
24/04/2025
Beam Dynamics will make its MPTS debut this year, showcasing its all-in-one platform designed to streamline workflows across live production, OB, broadcast, and...
24/04/2025
Test & measurement innovator, Leader Electronics of Europe, has announced that Czechia-based Comprimato, a leader in high performance software encoding and tran...
24/04/2025
CueScript, the leading developer of professional teleprompting solutions, took home two awards from NAB 2025 for its simple yet revolutionary solution that stre...
24/04/2025
Cobalt Digital, the leading designer and manufacturer of award-winning signal processing products and a founding partner in the openGear initiative, added anot...
24/04/2025
Dot Group, European specialists in IBM technologies for the media and entertainment industry, announces its participation at the Media Production & Technology S...
24/04/2025
Scality, a global leader in cyber-resilient storage for the AI era, today unveiled a first-of-its-kind unified software appliance developed in collaboration wit...
24/04/2025
Expanded applications for precise timing measurement across broadcast and professional AV sectors
Hitomi Broadcast, the market leader in audio/video alignment ...
24/04/2025
The Trump administration's efforts to shut down the Voice of America broadcast network were handed a setback this week when a federal court judge ordered th...
24/04/2025
BOSTON and NEW YORK The sports network NESN has announced that it has selected ViewLift provide solutions and technologies to upgrade the viewing and streaming ...
24/04/2025
YouTube is celebrating its 20th birthday with stats showing how popular the platform has become and by announcing some new features, including the ability for u...
24/04/2025
Warner Bros. Discoverys Max streaming service has announced a new U.S. product update introducing an Extra Member Add-On feature....
24/04/2025
Berklee in Puerto Rico Program to Host 30th Anniversary Concert The event will feature Grammy- and Latin Grammy-winning artist Miguel Zen n BM '98, cuatro...
24/04/2025
Detach Yourself from the Herd': Fito P ez Inspires at Berklee In a session hosted by the Mediterranean Music Institute, the influential Argentine musician...
24/04/2025
April 24th, 2025 Press Materials Available Here
Tribeca Festival 2025 Unveils Games Program
Featuring World Premieres and Playable Demos of Official Games Se...
24/04/2025
April 24th, 2025 Press Materials Available Here
Tribeca Festival 2025 Announce...
24/04/2025
April 24th, 2025 Press Materials Available Here
Tribeca Festival 2025 Announces Audio Storytelling Lineup
The Audio Storytelling Program To Kick Off With A L...
24/04/2025
Live From NFL Draft 2025: Van Wagner Turns Titletown Into Center of Fan Entertai...
24/04/2025
Live From the NFL Draft: NFL Media Splits Massive Production Between Green Bay a...
24/04/2025
Comcast Fills SpinCo Sports Executive Ranks With Matt Hong as President, Jeff Be...
24/04/2025
Back to All News
Crafting a Killer Goodbye to YOU'
The cast of YOU' pictured with fans outside the Mooney's pop-up in New York City. (Credit: Ne...
24/04/2025
Back to All News
You Read That Right: We're Introducing a New Way to Experience Subtitles
Product
24 April 2025
Global
Link copied to clipboard
Fifty ...
24/04/2025
Let me take you behind the scenes of a live sports production that came this close to not making it on air.
It was a quiet Monday-until I got a call from one ...
24/04/2025
Advancing AI requires a full-stack approach, with a powerful foundation of computing infrastructure - including accelerated processors and networking technologi...
24/04/2025
Get the controllers ready and clear the calendar - it's a jam-packed GFN Thu...
23/04/2025
WBD's Max viewership climbs 6%, boosted by The White Lotus' and The Pi...
23/04/2025
March shows a further downward trend of time spent watching television; following the February period when winter holidays contributed to a shorter time spent. ...
23/04/2025
During March, audiences in Mexico increased their streaming usage by 2.1 points compared to the previous month, accounting for 24.4% of TV viewing.
Disclaimer:...
23/04/2025
AANHPI audiences over index the total U.S. for share of time spent with Netflix ...
23/04/2025
SAN JOSE, Calif. Roku has announced new TVs, new streaming devices and significant upgrades to its user interface and software platforms that are designed to st...
23/04/2025
COW Job Listing: Opportunity for a Passionate Feature Film Editor - London-Based...
23/04/2025
RM Equity Partners Acquires MAGIX Software, Appoints Robert Rutkowski as CEO to ...
23/04/2025
All Men Are Wicked Western Shot with Blackmagic Design
Brie Clayton April 23, 2025
0 Comments
Blackmagic Pocket Cinema Camera 4Ks were put to the test...
Advancing AI requires a full-stack approach, with a powerful foundation of computing infrastructure - including accelerated processors and networking technologies - connected to optimized compilers, algorithms and applications.
