Posts in category: nVidia AI

There are more AI startups in healthcare than any other single industry. The number of AI startups in media and entertainment is about the same as that in retail. More than one in 10 of all AI startups is based in California.

How do we know this? NVIDIA Inception, our acceleration platform for AI startups, has now surpassed 8,500 members. That’s about two-thirds of the total number of AI startups worldwide, as estimated by Pitchbook. With total cumulative funding of over $60 billion and members in 90 countries, NVIDIA Inception is one of the largest AI startup ecosystems in the world.

With this type of scale, NVIDIA Inception is more than a singular program; it’s a reflection of the larger startup landscape. And there’s plenty that can be inferred based on this.

Data Across 8,500+ Startups

NVIDIA Inception figures show the United States leads the world in terms of both the number of AI startups, representing nearly 27 percent, and the amount of secured funding, accounting for over $27 billion in cumulative funding.

Of U.S.-based startups, 42 percent were based in California — more than one in 10 AI startups is based in the state — with 29 percent in the San Francisco Bay Area. This underscores the continued draw of the region for startup founders and VC funding.

Following the U.S. is China, in terms of both funding and company stage, with 12 percent of NVIDIA Inception members based there. India comes in third at 7 percent, with the United Kingdom right behind at 6 percent.

Taken together, AI startups based in the U.S., China, India and the U.K. account for just over half of all startups in NVIDIA Inception. Following in order after these are Germany, Russia, France, Sweden, Netherlands, Korea and Japan.

In terms of industries, healthcare, IT services, intelligent video analytics (IVA), media and entertainment (M&E) and robotics are the top five in NVIDIA Inception. AI startups in healthcare account for 16 percent of Inception members, followed by those in IT services at 15 percent. AI startups in IVA make up 8 percent, with M&E and robotics AI startups tied at 7 percent.

Details Spanning 3,000+ Startups Since 2020

More than 3,000 AI startups have joined NVIDIA Inception since 2020. Similar to data across Inception as a whole, AI startups from the U.S. account for the largest segment (27 percent), followed by China (12 percent), and India and the U.K. (tied at 6 percent).

Additionally, startups that have joined since 2020 are concentrated in the same top five industries, though in slightly different order. IT services leads the way at 17 percent, followed by healthcare at 16 percent, M&E at 9 percent, IVA at 8 percent and robotics at 5 percent.

Within the top two industries —  healthcare and IT services — there’s more detail among AI startups who have joined since 2020. The dominant segment within IT services is computer vision at 27 percent, with predictive analytics in second place at 9 percent. The top two segments in healthcare are medical analytics at 38 percent and medical imaging at 36 percent, though the fastest growth is among AI startups in the pharma and AI biology industries at 15 percent.

Virtual and augmented reality startup companies are far outpacing any other segment within M&E, mostly due to the pandemic. These startups are coming to NVIDIA Inception with a shared vision of building an ecosystem for the metaverse.

Disruption Through Startups 

Since Inception’s launch in 2016, it has grown more than tenfold. This growth has accelerated year over year, with membership increasing to 26 percent in 2020, and already reaching 17 percent in the first half of 2021.

NVIDIA Inception is a program built to accommodate and nurture every startup that is accelerating computing, at every stage in their journey. All program benefits are free of charge — there are no fees ever. And unlike other accelerators or incubators, startups never have to give up equity to join.

Startups are the single best lens into the future of modern AI, so join with us today by applying for NVIDIA Inception.

Deathtrap Dungeon: The Golden Room is a gripping choose-your-own-adventure story, but it’s no page-turner.

Based on the best-selling book of the same name, it’s an interactive film in which viewers become the player on their quest to find The Golden Room while facing down dungeon masters and avoiding traps.

NVIDIA RTX technology powers the real-time graphics and virtual sets behind this latest adaptation, which showcases the future of interactive storytelling on a virtual production stage.

On-Set Facilities (OSF) provided the technology for the virtual production. Using its own low-latency computing platform, the GODBOX powered by NVIDIA RTX, OSF enhanced virtual production workflows and delivered real-time compositing and previsualization for the interactive experience.

Bringing Virtual Sets to Life with NVIDIA RTX

When it comes to bringing VFX on set, OSF faced a common challenge — finding computers that could be configured for their creative teams and production needs. So they created their own on-set computer platform, GODBOX Workstations and Servers. It’s a synchronized real-time virtual production platform for low-latency, frame-accurate, virtual production applications and workflows.

From LED and in-camera VFX to mixed reality and motion capture, the GODBOX provides all the tools, features and solutions needed to set up and run a virtual production from any set.

For Deathtrap Dungeon, OSF laid the foundations during preproduction and previsualization. The team used virtual sets and real locations, and combined that with real-time visual effects to bring sets to life. Digitally creating the previsual assets allowed the team to specify the size of stages, amounts of props and how many physical sets were needed.

“The objective was to previsualize the final VFX on the set, so that the directors, actors and crew could all see the virtual world,” said Asa Bailey, director of virtual production at OSF. “The GODBOX delivers in-camera VFX and real-time compositing pipelines powered by NVIDIA RTX. The platform was specifically designed to work with all kinds of virtual productions.”

Throughout the preproduction and the film shoot, OSF used Unreal Engine to conduct virtual scouting sessions using the GODBOX cloud production VPN. Using the secure cloud platform, OSF tested the virtual sets and worked with the production to set lighting and camera movements, all before going on set.

With the RTX-powered GODBOX, OSF also delivered real-time compositing so the cast and crew can see their performance with the virtual set and characters.  The team combined green screen live action with virtual sets and VFX elements in Unreal Engine. Then they’d take the real-time composition and feed it through to large screens and projectors on set.

OSF’s GODBOX stays updated with the latest NVIDIA drivers, as well as recently released optimizations. This helps increase the stability of the machine, which becomes crucial when the cameras are rolling.

Learn more about On-Set Facilities, GODBOX low-latency computing and virtual production. And see other NVIDIA solutions in media and entertainment.

This GFN Thursday shines a spotlight on the latest games joining the collection of over 1,000 titles in the GeForce NOW library from the many publishers that have opted in to stream their games on our open cloud-gaming service.

Members can look forward to 14 games — including Evil Genius 2: World Domination from Rebellion and Escape From Naraka, which features RTX for Founders and Priority members — joining the GeForce NOW library this week. Team17’s Hell Let Loose, already streaming on GeForce NOW, has left early access with the introduction of the Soviet forces on the eastern front.

Getting Rebellious

Ready to take over the world? This GFN Thursday, stream several exciting titles from Rebellion — the award-winning British independent studio. Games include Evil Genius 2: World Domination, Evil Genius, Battlezone: Combat Commander and Zombie Army 4: Dead War.

Be the best bad guy you can be in Evil Genius 2: World Domination (Steam). Construct an evil lair, train your minions to carry out nefarious plans and fight against the Forces of Justice with an array of traps to achieve global domination in this satirical spy-fi lair-builder game.

Members can also look forward to the original Evil Genius coming to the cloud this week, as well as Battlezone: Combat Commander for space-shooting intergalactic battles. And gamers can sink their teeth into Zombie Army 4: Dead War for the fight against the undead Armageddon. With all of these new additions, more gamers now can play these awesome titles from Rebellion.

“GeForce NOW gives us a great opportunity to provide access to our catalog of games to even more gamers,” said Matt Jeffery, chief strategy officer at Rebellion. “NVIDIA makes onboarding games to the cloud a simple process, helping us bring our games to players on low-powered or incompatible devices, with the power of a real gaming rig.”

These newest additions join a host of other Rebellion titles available to stream on GeForce NOW, including Sniper Elite 3, Sniper Elite 4 and Sniper Elite V2 Remastered for players who aim to have a good time with a stealth story. Plus more undead action in the Zombie Army Trilogy and an adventure full of puzzles, traps and mummies in Strange Brigade that’s too good to keep under wraps.

Heating Things Up

The front line is calling. Hell Let Loose (Steam) from Team17 has left early access and is streaming in open access on GeForce NOW.

Charge into battle in this hardcore World War II first-person shooter with epic battles and players filling roles of infantry, tanks, artillery and the dynamically shifting front line. Battle across more than nine maps modeled on real-life locales. Fight in iconic battles from the Western Front, including Carentan, Omaha Beach, Foy and more. Members can experience the chaos of all-out war and stream Hell Let Loose with GeForce NOW this week.

Members can also check out other popular titles from Team17 in the GeForce NOW library, including multiplayer games like Overcooked! and Overcooked! 2 and Golf with Friends for gamers looking to play with friends. Or experience a variety of RPGs with the cozy and creative Hokko Life, the pirate adventure King of the Seas, and the alien world of Planet Alpha.

Up This Week

As always, GFN Thursday means more games. This week, members can journey to a nightmarish temple in Escape from Naraka and save their beloved from an evil demon of legend.

Complete challenges, unlock the secrets of the temple and use unique abilities to confront terrifying enemies and escape the temple. Founders and Priority members can play Escape from Naraka on GeForce NOW with beautifully ray-traced reflections, ray-traced shadows and RTX Global Illumination.

Here’s the complete list of 14 titles coming to the cloud this week:

• Escape from Naraka (day-and-date release on Steam, July 29)
• Starbase (day-and-date release on Steam, July 29)
• Banners of Ruin (day-and-date release on Steam, July 30)
• BUSTAFELLOWS (day-and-date release on Steam, July 30)
• Vesper (day-and-date release on Steam, July 30)
• Alan Wake (Epic Games Store)
• Battlezone: Combat Commander (Steam)
• Evil Genius (Steam)
• Evil Genius 2: World Domination (Steam)
• Lumberjack’s Dynasty (Epic Games Store)
• Moonlighter (Epic Games Store)
• Trials Fusion (Ubisoft Connect)
• X3: Farnham’s Legacy (Steam)
• Zombie Army 4: Dead War (Steam and Epic Games Store)

And while you decide which of these games to spend your weekend with, we have an important question for you.

who’s the most iconic villain in video game history?

— NVIDIA GeForce NOW (@NVIDIAGFN) July 28, 2021

Shout it out on Twitter, and we’ll see you next week!

9 Down, 14 letters: Someone skilled in creating and solving crossword puzzles.

This April, the fastest “cruciverbalist” at the ​​American Crossword Puzzle Tournament was Dr.Fill, a crossword puzzle-solving AI program created by Matt Ginsberg.

Dr.Fill perfectly solved the championship puzzle in 49 seconds. The first human champion, Tyler Hinman, filled the 15×15 crossword in exactly three minutes.

Though Ginsberg has published crossword puzzles for the New York Times, he has trouble solving puzzles, even his own. After attending a crossword tournament over a decade ago, Ginsberg decided to create a crossword-solving program to compete against top-tier word nerds.

Ginsberg spoke with NVIDIA AI Podcast host Noah Kravitz about his decade-long journey creating Dr.Fill and where he envisions it going in the future.

Key Points From This Episode:

• Ginsberg partnered with UC Berkeley’s natural language processing team. By combining their code bases, Dr.Fill outperformed the human competitors at this year’s tournament.
• Dr.Fill performs better on unthemed rather than themed puzzles. This is where the natural language group comes into play, helping Dr.Fill interpret clues and difficult crossword themes.

Tweetables:

“I’m looking forward to the crossword tournament next year because I know we’re going to be working hard to make the program better.” — Matt Ginsberg [11:02]

You Might Also Like:

Otter.ai CEO Sam Liang on Bringing Live Captions to a Meeting Near You

Sam Liang is making things easier for the creators of the NVIDIA AI Podcast — and just about every remote worker. He’s the CEO and co-founder of Otter.ai, which uses AI to produce speech-to-text transcriptions in real time or from recording uploads.

How Deep Learning Can Translate American Sign Language

Rochester Institute of Technology computer engineering major Syed Ahmed, a research assistant at the National Technical Institute for the Deaf, uses AI to translate between American Sign Language and English. Ahmed trained his algorithm on 1,700 sign language videos.

Pod Squad: Descript Uses AI to Make Managing Podcasts Quicker, Easier

Serial entrepreneur Andrew Mason is making podcast editing easier and more collaborative with his company, Descript Podcast Studio, which uses AI, natural language processing and automatic speech synthesis.

Subscribe to the AI Podcast

Get the AI Podcast through iTunes, Google Podcasts, Google Play, Castbox, DoggCatcher, Overcast, PlayerFM, Pocket Casts, Podbay, PodBean, PodCruncher, PodKicker, Soundcloud, Spotify, Stitcher and TuneIn. If your favorite isn’t listed here, drop us a note.

Make the AI Podcast Better

Have a few minutes to spare? Fill out this listener survey. Your answers will help us make a better podcast.

Ever see a virtual kitchen materialize in real-time? If you caught NVIDIA CEO Jensen Huang’s keynote for our March 2021 GPU Technology Conference you’re no doubt wondering about more than a few of the presentation’s magic tricks.

With the premiere of “Connecting in the Metaverse: The Making of the GTC Keynote,” Wednesday, Aug. 11, at 11 a.m. Pacific time, NVIDIA team members will reveal the story behind the story told at GTC.

Designed to entertain and inform, GTC keynotes are always filled with cutting-edge demos highlighting NVIDIA’s advancements in supercomputing, deep learning and graphics.

This year, however, with NVIDIA’s team working remotely to create a presentation for attendees flocking to an entirely virtual gathering, those technologies were crucial for making the keynote itself.

The highlight: the reveal of Huang’s virtual kitchen, complete with a digital clone of the man himself.

The half-hour documentary film debuting Wednesday tells the tale of the small team of remote artists who were able to blur the line between real and rendered on a tight deadline with NVIDIA Omniverse, a platform for connecting 3D worlds into a shared virtual space.

Go to our SIGGRAPH 2021 landing page to watch the video, and browse all our events at this year’s conference.

Immediately after the video’s debut, those who are registered for SIGGRAPH can join the team at NVIDIA who will dig into the story told in the video for a live Q&A at 11:30 a.m. Pacific time.

To participate in the panel, register for SIGGRAPH using the code “NVIDIA21” for a free basic pass or $25 off an enhanced or ultimate pass.

You’ll be able to find all the details you’ll need to join the panel once they’re posted on SIGGRAPH’s website.

This GFN Thursday brings in hordes of fun — and a whole lot of orcs. Orcs Must Die! 3, the newest title from the action-packed, orc-slaying series from Robot Entertainment, is joining the GeForce NOW library when it releases tomorrow, Friday, July 23.  

In addition, 10 more games are coming to the service this week.

Play Your Games, Your Way

Gaming on GeForce NOW means having instant access to over 1,000 PC games streaming from the cloud. Whether it’s a low-powered PC, Macs, Chromebooks, SHIELD TVs or Android and iOS mobile devices, GeForce NOW supported devices play real PC games with GeForce levels of performance.

And you don’t just get to play real PC versions of games optimized on GeForce NOW compatible devices. The games you play are the ones you own, streaming from popular stores like Steam, Epic Games Store, Ubisoft Connect and GOG.COM. With GeForce NOW, you never have to rebuy a different version of that game to play it on multiple compatible devices.

Members can play awesome PC games, like Orcs Must Die! 3 in high-powered detail and battle across all of their GeForce NOW compatible devices.

Bring the Mayhem

Orcs Must Die! 3 (Steam) challenges players to slice, burn, toss, zap, grind and give it all they’ve got to keep massive hordes of orcs at bay across the battlefield in an effort to secure the castle walls and victory. Battle with a buddy or slay your enemies solo against enormous orc armies with weapons and traps of your choice.

Players can experience all-new war scenarios that pit players against overwhelming legions of orcs. But don’t worry, players have an arsenal of magic, weapons and new War Machines: traps on an oversized scale that range from mega flip traps that launch orcs ragdolling off the castle walls to mega boom barrel launchers that unleash pyrotechnic glory.

Enjoy a new and exciting story set 20 years after the previous game with fresh characters. Play for glory in weekly challenges to see how long you can survive in Endless Mode to put your name on the leaderboard. Then, survive Scramble Mode and face off against evolving orcs with sinister surprises. Finally, take things a step further in the Drastic Steps campaign as the mayhem takes to the skies against flying orcs.

With Orcs Must Die! 3 releasing on Steam and joining the GeForce NOW library, more gamers than ever before can experience the thrill of stomping orcish hordes — regardless of their low-powered rigs.

“We love that our players can experience all of the action and all of the orcs on any GeForce NOW compatible device,” said Patrick Hudson, CEO at developer Robot Entertainment. “It’s great that players who may not have powerful devices can still play the real PC version of our game at orc-slaying power.”

Members can get ready to slay and play Orcs Must Die! 3 on GeForce NOW when it releases tomorrow, July 23.

Game Time

Orcs Must Die! 3 isn’t the only new game this GFN Thursday. Members can look out for these sweet titles ready to stream this week:

• Cris Tales (day-and-date release on Steam and on Epic Games Store, July 20)
• Death’s Door (day-and-date release on Steam, July 20)
• Warhammer 40,000: Battlesector (day-and-date release on Steam and Epic Games Store, July 22)
• Orcs Must Die! 3 (day-and-date release on Steam, July 23)
• Alan Wake’s American Nightmare (Epic Games Store)
• Arcana Heart 3 LOVEMAX SIXSTARS!!!!!! XTEND (Steam)
• Disgaea 2 PC (Steam)
• Edge of Eternity (Epic Games Store)
• Pathfinder: Kingmaker – Enhanced Plus Edition (Epic Games Store)
• Siege Survival: Gloria Victis (Epic Games Store)
• World of Warplanes (Steam)

What games are you playing as we charge into the weekend? Let us know on Twitter or in the comments below.

Walk into a store. Grab your stuff. And walk right out again, without stopping to check out.

In just the past three months, California-based AiFi has helped Choice Market increase sales at one of its Denver stores by 20 percent among customers who opted to skip the checkout line.

It allowed Żabka, a Polish convenience store chain, to provide faster checkout for morning train commuters.

It helped pro-racing team Penske and Verizon run a dinky 200-square-foot store at the Indy500, so race fans could quickly get back to the action.

And on Wednesday AiFi announced an expanded partnership with Loop Neighborhood to introduce its computer vision, camera-only platform into stores in California, starting with two Bay Area locations.

AiFi, a member of the NVIDIA Inception accelerator for AI and deep learning startups, has moved out of the proof-of-concept stage and into stores across the world.

Its technology makes shopping more convenient and helps retailers better understand their customers.

“Retailers can now get as much information about physical shopping habits as online stores are getting from ecommerce,” said AiFi CEO and co-founder Steve Gu, a veteran of Apple and Google.

AiFi’s ability to analyze shopper habits is even more impressive because its stores don’t need to buy costly sensors and RFID tags.

Instead, the company uses real-time image recognition and edge AI powered by NVIDIA GPUs to recognize the items shoppers select and charge them, usually through an app linked to the customer’s credit card.

It’s not an easy task in busy stores stocked with many hundreds of items, but the five-year-old company’s technology now achieves an accuracy rate of 99 percent.

To date, more than 15 stores worldwide are putting the company’s technology to work. Those stores are already serving satisfied customers who return again and again.

At Choice Market, 60 percent of shoppers who tried the checkout-free option used it again within a month. Twenty percent came back three times.

The computer-vision-powered system uses NVIDIA Metropolis for smart video analytics. It works smoothly alongside the store’s traditional checkout system, and is integrated with the Choice Now app, where customers can shop checkout-free, and place online orders and arrange pickups.

AiFi is revolutionizing retail operations with AI. At the Indy500 auto race earlier this year, Penske Entertainment’s nano-store allowed fans to buy snacks, beverages and merchandise with an app. No need to stop and swipe a credit card.

This speed translates well to any kind of store where people need to slip in and out in a hurry. Żabka partnered with AiFi to open its first public autonomous store in Poznan, Poland, quickly drawing huge amounts of foot traffic from commuters going to and from a nearby train station.

With AiFi integrated with the Żappka App, which has over 5 million users, harried commuters could hustle out the door with a newspaper and a morning coffee.

More stores equipped with AiFi’s technology are coming. In France, the company is working with Carrefour on what the retailer calls its “Flash LabStore,” a frictionless store at its headquarters in Massy.

And in the UK, Britain’s fourth-largest supermarket, Morrisons, is working with AiFi to test out a store with no checkout.

These stores represent a growing number of collaborations with major retailers that see AiFi’s combination of AI and computer vision as the key to a brick and mortar retail renaissance that will ultimately put more goods in front of more customers, in more places.

Learn more about NVIDIA’s AI platform powering intelligent retail stores. 

The final push for the hat trick came down to the wire.

Five minutes before the deadline, the team submitted work in its third and hardest data science competition of the year in recommendation systems. Called RecSys, it’s a relatively new branch of computer science that’s spawned one of the most widely used applications in machine learning, one that helps millions find what they want to watch, buy and play.

The team’s combination of six AI models packed into the contest’s limit of 20 gigabytes all of the smarts it culled from studying 750 million data points. An unusual rule in the competition said the models had to run in less than 24 hours on a single core in a cloud CPU.

They hit the submission button and waited.

Twenty-three hours and 40 minutes later an email arrived: They hit No. 1 on the leaderboard.

Right Under the Buzzer

On June 28 it was official, a seven-member NVIDIA team won for the second time the ACM RecSys Challenge.

“The email came in right under the buzzer — 20 minutes later and we would have timed out,” said Chris Deotte, one of several team members who’s also a grandmaster in Kaggle competitions, the online Olympics of data science.

“We were really on the edge,” said Benedikt Schifferer, a teammate who helps design NVIDIA Merlin, a framework to help users quickly build their own recommendation systems.

GPUs could have busted through the inference job in a fraction of the time. Adapting the work to one CPU core “was like going back to the distant past,” said Gilberto “Giba” Titericz, a Brazil-based Kaggle grandmaster on the team.

In fact, once the competition was over, the team demonstrated the inference job that took nearly 24 hours on a CPU core could run on a single NVIDIA A100 Tensor Core GPU in just five and a half minutes.

Sorting 40M Items a Day

For that competition, Twitter gave participants millions of data points a day for 28 days and asked them to predict which tweets users would like or retweet. It was an industrial-strength challenge from the leading technical conference on RecSys, an event that draws a who’s who of top engineers from Facebook, Google, Spotify and other players.

The discipline is as hard as it is helpful. Recommendation systems fuel our digital economy, serving up suggestions faster and smarter than a traditional search.

Industry challenges help advance the field for everyone, whether they’re seeking the perfect gift for a spouse or trying to find an old friend online.

Three Wins in Five Months

Earlier this year, the full NVIDIA team led a field of 40 in the Booking.com Challenge. They used millions of anonymized data points to correctly predict the final city a vacationer in Europe would choose to visit.

In June, another top recsys contest, the SIGIR eCommerce Data Challenge, set an even higher hurdle.

The annual meeting of the Special Interest Group on Information Retrieval, SIGIR, draws experts from companies that span Alibaba to Walmart Labs. Its 2021 challenge provided 37 million data points from online shopping sessions and asked participants to predict which products users would buy.

Overlap with the ACM contest forced the NVIDIA team to split into two groups that coordinated their efforts between the contests. Ratcheting up the pressure, some team members were heads down writing a paper for the ACM RecSys conference.

The Art of the Fast Break

Two factors propelled a five-person NVIDIA team with members spread across Brazil, Canada, France and the U.S. to the best overall performance, taking first or second place in every leaderboard. They made a big bet on Transformer models developed for natural-language processing and increasingly adopted for recsys, and they understood the art of the handoff.

“As one member is going to bed another picks up the work in a different time zone,” said Even Oldridge, who leads the Merlin group.

“When it all clicks, it’s very effective, and I’m amazed at what we’ve accomplished in the last year building our internal knowledge and our standing in the recsys community to the point where we could win three major competitions in five months,” he said.

Respecting User Privacy

The contest required models to make predictions with no background on users beyond their current browsing session.

“That’s an important task because sometimes users want to browse anonymously, and some privacy laws limit access to historical information,” said Gabriel Moreira, a senior Merlin researcher in São Paulo who led NVIDIA’s SIGIR team.

The competition marked the first time the team used only Transformer models in their solution to a challenge. Moreira’s team aims to make the massive neural networks more easily available to every Merlin customer.

From a Hat Trick to a Haul

On June 30, we notched a fourth consecutive win in RecSys, what hockey players call a haul. MLPerf, an industry benchmarking group, announced that NVIDIA and its partners set records in all its latest training benchmarks, including one in recommendation systems.

The team behind that effort described its work training a recommendation system in less than a minute on 14 NVIDIA DGX systems, a 3.3x speedup compared to their submission a year ago.

Sharing Lessons Learned

The competitions fuel ideas for new techniques that find their way into recsys frameworks like Merlin and related tools, papers and online classes held by the NVIDIA Deep Learning Institute. The ultimate goal: Help everyone succeed.

In interviews NVIDIA’s recsys experts freely shared their know-how — part art, part science.

A Pro Tip on RecSys

One best practice is using a diversity of models that work together as an ensemble.

In the ACM RecSys Challenge, the team used both tree and neural-network models. The outputs from one stage became inputs for the next in a process called stacking.

“A single model can make a mistake due to a data error or convergence issue, but if you take an ensemble of several models, it’s very powerful,” said Bo Liu, the newest member of NVIDIA’s Kaggle grandmaster team.

Meet RecSys Experts Online

On July 29, you can meet RecSys experts from Facebook, NVIDIA and TensorFlow to learn more about how to create great recommender systems.

Top game artists, producers, developers and designers are coming together this week for the annual Game Developers Conference. As they exchange ideas, educate and inspire each other, the NVIDIA Studio ecosystem of RTX-accelerated apps, hardware and drivers is helping advance their craft.

GDC 2021 marks a major leap in game development with NVIDIA RTX technology integrated in the latest releases of Unity, Unreal Engine 4, Toolbag and more. They’re supported by the July NVIDIA Studio Driver, available today, providing peak performance and reliability.

Developers can create better looking games, in less time, without worrying about their systems crashing, with NVIDIA Studio.

Game Development Runs Faster with NVIDIA Studio

The two most popular PC game engines, Unity and Unreal Engine, recently received additional RTX benefits.

The new 2021.2 beta release of Unity delivered native support for NVIDIA DLSS, allowing game developers to easily incorporate advanced AI rendering into their games. DLSS produces image quality that’s comparable to native resolution — and sometimes even better — while only conventionally rendering a fraction of the pixels, boosting real-time performance for more engaging experiences and saving artists valuable exporting time.

DLSS SDK 2.2.1, the latest offered by NVIDIA and built into Unity 2021.2, brings a new blueprint function to enable the optimal image quality for a particular resolution, called “Auto” mode. There’s also an optional sharpening slider so developers can further tune their visuals.

Unreal Engine 4.27, currently in preview, included an experimental feature called Eye-Tracked Foveated Rendering. The technique renders a single image at varying resolutions, sharpening the point of focus, while blurring other parts, to mimic human eyesight.

It’s perfect for extended reality with improved performance on NVIDIA RTX GPUs, using NVIDIA Variable Rate Shading, and no discernable loss of picture quality. In addition, GPU Lightmass baking built on RTX ray tracing introduced parameters to better control lighting and levels of detail in production assets.

Marmoset Toolbag 4.03 sports a new ray-tracing engine, optimized to run on all modern GPUs. Even faster ray-traced results are achieved with native hardware support of NVIDIA RTX devices.

The most recent update added RTX-accelerated AI denoising, allowing game artists to quickly visualize materials with photorealistic lighting and shadows.

RTX-accelerated ray racing with improved shading, global illumination and reflections raise the visual quality bar, while RTX-accelerated baking speeds up asset creation.

NVIDIA Omniverse is a platform for 3D content creation and collaboration. It was built from the ground up to be easily extensible and customizable with a modular development framework. The platform includes ready-made Omniverse Apps like Machinima and Audio2Face, plus a collection of over 200 Omniverse Kit Extensions, small pieces of code purpose-built to achieve a specific task.

Game developers can use the prebuilt apps or extensions, or easily build their own tools on Omniverse Kit, a robust system allowing coders with basic programming knowledge to build extensions, apps and microservices to assist in content creation pipelines.

Developers can learn more about Omniverse in the GDC session Collaborative Game Development with NVIDIA Omniverse, taking place from 8:30-9:30 a.m. PT on July 22. The session will feature tips on collaborative workflows between leading industry applications such as Unreal Engine 4, 3ds Max, and Maya, plus an introduction on how to build on Omniverse Kit. Interested developers can register here.

The Studio Advantage, Built for the Bold

NVIDIA Studio ushered in a new era of creative performance with laptops and desktops purpose-built to power the world’s most innovative minds. Packed with industry-leading RTX GPUs, these machines deliver unprecedented levels of computing power.

Future game developers and content creators can unleash their creativity and build magnificent worlds with the latest RTX 30-Series GPU-powered NVIDIA Studio laptops.

Perfect for students heading back to school, Studio laptops accelerate more than just the latest game engines, they power dozens of applications in STEM — including engineering, computer science, data science and economics applications — plus the apps creators rely on. The latest selection of Studio laptops can be found in the Studio Shop.

Together with game engine and creative app developers, teams of testers and engineers are continually optimizing the way NVIDIA hardware works with top software — enhancing features, reducing the repetitive and speeding up workflows. Studio Drivers undergo extensive testing to deliver the performance and reliability developers need, helping them create the blockbuster games at the speed of imagination.

Further Boost Creativity With the July Studio Driver

The July NVIDIA Studio Driver available today features support for updates to Unity, Unreal Engine, Toolbag, Omniverse and more.

Enscape 3.1, dropping July 21, adds a new NVIDIA real-time denoiser and support for NVIDIA DLSS, designed for real-time engines utilizing NVIDIA RTX GPUs.

This enables smoother viewport visibility, as well as the ability to render at lower resolutions, enabling higher framerates, using AI super resolution to upscale the image to equal if not higher visual fidelity.

Pixar Animation Studios RenderMan 24 added RenderMan XPU, a look-development focused GPU-accelerated ray tracer.

Together with AI denoising in the viewport, RenderMan XPU enables artists to interactively create their art and view an image that is predictive of the final frame render.

Click here to download the Maya teapot asset used in performance testing.

Topaz Video Enhance AI now offers Slow Motion, a new RTX GPU Tensor Core powered AI feature that generates a high-quality, smooth, slow-motion capture with minimal artifacts.

Crucially, eliminating the need for an expensive high-frame-rate camera.

Finally, gamers and content creators who use Discord to collaborate and share content with friends can use the new NVDEC integration, exclusive to NVIDIA GPUs, for accelerated video decoding. This lets them share screens and stream over Discord with reduced resources for video and results in better gaming performance.

Stay up to date on new Studio products by subscribing to the NVIDIA Studio newsletter and following us on Facebook, Twitter and Instagram.

A pair of new demos running GeForce RTX technologies on the Arm platform unveiled by NVIDIA today show how advanced graphics can be extended to a broader, more power-efficient set of devices.

The two demos, shown at this week’s Game Developers Conference, included Wolfenstein: Youngblood from Bethesda Softworks and MachineGames, as well as The Bistro from the Open Research Content Archive running in real time on a MediaTek Arm platform with ray-traced graphics.

RTX has redefined the industry. We’re now investing in new platforms where we can deploy advanced graphics so gamers have more choice. The performance and energy efficiency of ARM CPUs with NVIDIA technologies can open an entirely new class of PCs.

“RTX is the most groundbreaking technology to come to PC gaming in the last two decades,” said PC Tseng, general manager of MediaTek’s Intelligent Multimedia Business Unit.“MediaTek and NVIDIA are laying the foundation for a new category of Arm-based high-performance PCs.”

RTX on Arm in Action

Showing the potential for NVIDIA RTX on Arm, developer Machine Games packed the Wolfenstein: Youngblood demo with beautiful, ray-traced reflections, all accelerated by NVIDIA DLSS, which uses GPU-accelerated deep-learning algorithms to boost frame rates.

NVIDIA also showed how RTX can enhance the The Bistro demo, which portrays a detailed, ray-traced urban scene in France, while running on an Arm-based system.

Both were demonstrated on an NVIDIA GeForce RTX 3060 GPU paired with a MediaTek Kompanio 1200 Arm processor. Wolfenstein: Youngblood uses the idTech game engine made by id Software, while The Bistro uses NVIDIA’s sample framework.

View the demos here.

The demos are made possible by NVIDIA extending support for its software development kits for implementing five key NVIDIA RTX technologies to Arm and Linux.

They include:

• Deep Learning Super Sampling (DLSS), which uses AI to boost frame rates and generate beautiful, sharp images for games
• RTX Direct Illumination (RTXDI), which lets developers add dynamic lighting to their gaming environments
• RTX Global Illumination (RTXGI), which helps recreate the way light bounces around in real-world environments
• NVIDIA Real-Time Denoisers (NRD) a denoising library that’s designed to work with low ray per pixel signals
• RTX Memory Utility (RTXMU), which optimizes the way applications use graphics memory

The Potential for RTX on ARM

GeForce RTX technologies — including GPU-accelerated ray tracing, NVIDIA DLSS and other AI-powered innovations — have made a significant impact on real-time graphics since their introduction in 2018.

The world’s leading publishers have used NVIDIA RTX technologies to set apart their top franchises. RTX technologies are now available in an all-star list of gaming franchises, including Battlefield, Call of Duty, Cyberpunk, DEATH STRANDING, Doom, Final Fantasy, Fortnite, LEGO, Minecraft, Quake, Rainbow Six, Red Dead Redemption, Rust, Tomb Raider, Watch Dogs and Wolfenstein.

The news garnered widespread industry support.

• “NVIDIA extending RTX support to Arm and Linux has the potential to benefit games and industries such as automotive, where leading manufacturers use Unreal Engine not only for design visualization but also for digital cockpits and infotainment” said Nick Penwarden, vice president of engineering, Epic Games. “We always welcome powerful features and SDKs that can be leveraged across many platforms.”
• “Wolfenstein: Youngblood is the first RTX PC game to be shown on an Arm-based system, a testament to the flexibility, power and optimized nature of the iD Tech engine,” said Machinegames CTO Jim Kjellin. “An iD Tech-based game running on an Arm CPU with ray tracing enabled is a significant step in a journey that will result in many more gaming platforms being available to all game developers.”
• “RTX support for Arm and Linux opens up new opportunities for game developers to provide more immersive experiences on a wider variety of platforms,” said Mathieu Muller, senior technical product manager of high-end graphics at Unity. “With GeForce RTX’s cutting-edge graphics features, Unity developers targeting Arm platforms will have more tools in their toolbox to create with.”

The RTXDI, NRD and RTXMU SDKs for Arm with Linux and Chromium are available now. RTXGI and DLSS will be coming soon. For more information, contact NVIDIA’s developer relations team or visit developer.nvidia.com.