Every year, Nvidia goes through hundreds of startups that use their graphics processing units (GPUs) to accelerate non-graphics applications such as artificial intelligence or pattern recognition. And there are a handful who are worth an investment or attention with their Inception program.
I've been to Nvidia's GPU Technology Conference every year, but I will not be coming to the next week because it's overlapping with the Game Developers Conference in San Francisco. However, I had the opportunity to gather some information about the startups in Inception, Nvidia's virtual accelerator program, which allows GPU-based startups to be brought to market faster.
Through the program, Nvidia offers GPU discounts, technical guidance and connectivity to businesses in places like the Inception Showcase at GTC.
This year's game includes:
Shone brings autonomous cargo ships to the market in San Francisco. It works with cargo shipping to provide similar benefits as autonomous vehicle functions in the automotive industry. Shone applies GPUs to cargo ship data such as sonar, radar, GPS, and ship-to-ship tracking (AIS). This allows terabytes of training data to be processed quickly according to Shone's custom algorithms for developing sea freight vessel awareness, navigation, and control systems.
Maritime transport is the backbone of the world economy, but it is not sustainable. Shone equips freighters with autonomous technologies to increase safety and efficiency, and builds on existing sensors aboard cargo ships (radar, AIS, GPS, anemometers and others). It also adds some extra sensors, mostly cameras.
Shone guides the cameras through the cameras and performs a sensor fusion between the various sensing sensors to gain a probable understanding of the ship's environment. Then only the relevant information that is useful to the crew depending on the context is highlighted. Shone's smart copilot also recommends safe and efficient decisions to the crew.
The Goat Group has raised $ 100 million in new funds for its sneaker market, which AI uses to authenticate shoemakers for coveted collectibles. The startup in Culver City, California, has grown to more than 500 employees, attracts more than 10 million users, and has grown astronomically since its inception in 2015.
The authentication process uses machine learning and human expertise to build trust and transparency for the sneaker industry. With its innovative technology and innovative service, Goat Group aims to be the most reliable source for discovering, buying and selling authentic sneakers.
Lift has a platform that promotes the work of professional translators and enhances domain expertise for specific projects with its hybrid human-machine training pipeline. With San Francisco startup software, users can review and translate each line of text in one language, including whole lines of translation suggestions.
Lilt aims to provide high-quality translations that enable companies to bring their products to more countries faster than ever His KI works with human translators to improve quality and uses human domain experts to better serve customers.
This approach takes people into the loop to use machine translation platforms that combine neural networks with people's expertise Translators The system works like a bilingual text with predictive text, suggests translations, and learns and adapts to what the Accept or reject translators.
The company said it is dramatically reducing the cost and time of effective translation services, allowing more companies to expand into new markets and operate internationally. This opens new doors to the global knowledge-based economy and drives Lilt's vision of making information accessible to everyone, beyond a lingua franca.
Babblelabs is modernizing the digitized language with AI and will soon emerge as a major player in solving the problem of the cocktail party of background noise. The Campbell Clear Cloud API, California, trained on Nvidia GPUs, provides automatic speech enhancement and noise reduction for sound processing.
The company improves speech understanding for both humans and machines on platforms such as phones, cars, devices and voice monitors. It also has applications in industries ranging from Home Internet of Things (IoT) devices to customer service, training and law enforcement.
It offers "industry-best voice quality," clarity, and personalization through a unique combination of linguistics, deep learning algorithms, and a huge training body. The resulting audio, video, speech intelligibility and speech recognition in noisy environments mean better device performance, better quality of service, improved security and lower overall costs.
Babblelabs builds a unique technology and unique business in language, AI and optimization on computational implementations. It builds unique neural network structures that are trained and tuned by language experts to extract surprisingly clean speech from noise and to be implemented on GPUs, CPUs, DSPs, neural network accelerators, and microcontrollers.
This means that new speech enhancements are possible in speech recognition and speech analysis in a variety of popular devices for more convenience, accuracy and privacy. His target applications are mobile telephony, critical corporate communications, the automotive industry and the language-based Internet of Things.
Mapillary develops detailed maps through the integration of Computer Vision technology and Community Collaboration. Based in Malmö, Sweden, the startup will gather street-level images from cameras to create visualizations of the world, improve maps, help cities plan their development, and contribute to the development of the automotive industry.
The startup leverages collaboration and computer vision to meet the rapidly growing need for detailed and accurate maps. The combination of rapid advances in autonomous driving and urban development forces scalable updating of maps. This explains why the global digital card market will rise to over $ 8 billion by 2025.
Unlike other map data providers, Mapillary is independent, collaborative, and on-demand. All 465 million images on the Mapillary platform have been uploaded by individuals, organizations, and even countries.
The images and data are analyzed automatically and on a scale using Mapillary's Computer Vision technology. With a large and diverse range of road-level imagery from 190 countries, including cameras ranging from smartphones to professional rigs, we have been able to build some of the computer vision technologies for street scenes.
Customers include mapping companies such as Hier and Mapbox, automotive companies such as Toyota and AID (an Audi subsidiary focused on autonomous vehicles), and local authorities ranging from major cities such as Los Angeles to small towns such as Clovis, New Mexico. Mapillary has raised about $ 24.5 million from Atomico. Sequoia Capital, BMW i Ventures and Samsung Catalyst Fund. CEO Jan Erik Solem co-founded the company in 2013.
Prenav is a drone developer focused on infrastructure inspections. The startup in Redwood City, California, is using its home-grown eyelid and guidance system to bring drones close to industrial inspection points and allow them to navigate indoors without access to GPS. Prenav's inspections include US cell towers and energy infrastructure.
The company developed highly automated drones and deep-learning algorithms for the inspection of critical infrastructures (bridges, dams, power plants and electric towers). This is significant as the US infrastructure is aging and the US Civil Society of Civil Engineers estimates a funding gap of $ 2 trillion over the 10 years from 2016 to 2025.
This means ongoing inspections and maintenance become more important Structures are extended well beyond their intended lifetime. Currently, inspections of large concrete and steel structures are mainly performed by workers on ropes, scaffolding or bucket vehicles, and the overall process is slow, expensive, dangerous and inaccurate. As a result, critical damage and defects are sometimes overlooked, with catastrophic consequences such as the collapse of the Ponte Morandi Bridge in Italy, the crisis around the Oroville Dam and 17 forest fires in 2017 attributed to PG & E equipment by Cal Fire were.
Prenav's drones fly near structures (ie in GPS-denied environments, such as up close, under, or in confined spaces). The system uses a unique LIDAR-enabled ground-based base station for GPS-denied navigation and then takes hundreds of high-resolution photos that are stitched together to build a detailed 3D reconstruction or "digital twin" of the structure under test. Custom deep learning algorithms then look for common damage and defects, such as: Cracks in concrete, cracks in steel, missing rivets, rusted splinters, flooded insulators, etc.
Prenav can digitize large concrete and steel structures with 0.2 millimeter surface resolution. The system consists of a precisely guided drone for GPS-rejected environments and Prenav.xyz, a web-based platform for visualizing high-resolution digital twins and detecting damage.
The company raised $ 6.5 million in seeds in 2016 and has recently emerged from research and development with customers in the transportation, energy, telecommunications and construction sectors.
With Vyasa's deep learning platform, customers can conduct life-science queries on science, healthcare, marketing, law, and business intelligence. The Boston-based startup releases GPUs to help customers unlock data silos for insights.
The company provides a highly scalable platform for deep learning (cortex) for collaborative data science in project teams. With its library of analysis modules, Cortex enables novel, deep-learning analysis of image, text, and quantitative and small composite data sources for life science and healthcare applications. These include crystal morphology for formulation, screening of high-content cell assays, competitive intelligence, de novo compound design, and EHR analytics.
Kinetica provides a platform that combines streaming and historical data with location intelligence and machine learning analysis. Companies in the automotive, energy, telecoms, retail, and financial services industries use the platform's GPU-accelerated computing power to create custom analytics applications that deliver real-time results.
The analysis of Kinetica is constantly performed and updated in real time other systems based on the results. Active Analysis enables customers to build custom business decision-making applications based on historical information, machine learning, and predictive analytics AI.
The Kinetica Active Analytics Platform enables organizations to continuously and automatically combine, analyze, and act on billions of live, streaming, and historical data events. With Kinetica, a company can develop intelligent, analytical applications to enable digital business functions such as automated retail refills, adaptive telecoms coverage, continuous financial risk assessment, pattern recognition for AV activation, and predictive analytics in pharmaceutical research.
The Kinetica platform includes a distributed Nvidia GPU-accelerated in-memory database that uses a powerful combination of CPUs and GPUs to analyze massive, complex datasets with millisecond response times.
Kinetica has active analysis capabilities that dramatically simplify the architecture of large-scale, intelligent analytical applications that combine historical analysis, streaming analysis, graph analysis, location intelligence, and machine learning-driven analytics.