Monthly Archives: April 2022

In recent years, we have seen accelerated OLED materials development, aided by software tools based on machine learning and Artificial Intelligence. This is an excellent development which contributes to the continued improvement in OLED efficiency, brightness and lifetime.

Kyulux's Kyumatic AI material discover system

The promise of these new technologies is the ability to screen millions of possible molecules and systems quickly and efficiently. Materials scientists can then take the most promising candidates and perform real synthesis and experiments to confirm the operation in actual OLED devices.

The main drive behind the use of AI systems and mass simulations is to save the time that actual synthesis and testing of a single material can take - sometimes even months to complete the whole cycle. It is simply not viable to perform these experiments on a mass scale, even for large materials developers, let alone early stage startups.

In recent years we have seen several companies announcing that they have adopted such materials screening approaches. Cynora, for example, has an AI platform it calls GEM (Generative Exploration Model) which its materials experts use to develop new materials. Another company is US-based Kebotix, which has developed an AI-based molecular screening technology to identify novel blue OLED emitters, and it is now starting to test new emitters.

The first company to apply such an AI platform successfully was, to our knowledge, Japan-based Kyulux. Shortly after its establishment in 2015, the company licensed Harvard University's machine learning "Molecular Space Shuttle" system. The system has been assisting Kyulux's researchers to dramatically speed up their materials discovery process. The company reports that its development cycle has been reduced from many months to only 2 months, with higher process efficiencies as well.

Since 2016, Kyulux has been improving its AI platform, which is now called Kyumatic. Today, Kyumatic is a fully integrated materials informatics system that consists of a cloud-based quantum chemical calculation system, an AI-based prediction system, a device simulation system, and a data management system which includes experimental measurements and intellectual properties.

Kyulux is advancing fast with its TADF/HF material systems, and in October 2021 it announced that its green emitter system is getting close to commercialization and the company is now working closely with OLED makers, preparing for early adoption.

Continued here:

How machine learning and AI help find next-generation OLED materials - OLED-Info

An artificial intelligence tool that reads chest X-rays without oversight from a radiologist got regulatory clearance in the European Union last week a first for a fully autonomous medical imaging AI, the company, called Oxipit, said in a statement. Its a big milestone for AI and likely to be contentious, as radiologists have spent the last few years pushing back on efforts to fully automate parts of their job.

The tool, called ChestLink, scans chest X-rays and automatically sends patient reports on those that it sees as totally healthy, with no abnormalities. Any images that the tool flags as having a potential problem are sent to a radiologist for review. Most X-rays in primary care dont have any problems, so automating the process for those scans could cut down on radiologists workloads, the Oxipit said in informational materials.

The tech now has a CE mark certification in the EU, which signals that a device meets safety standards. The certification is similar to Food and Drug Administration (FDA) clearance in the United States, but they have slightly different metrics: a CE mark is less difficult to obtain, is quicker, and doesnt require as much evaluation as an FDA clearance. The FDA looks to see if a device is safe and effective and tends to ask for more information from device makers.

Oxipit spokesperson Mantas Miksys told The Verge that the company plans to file with the FDA as well.

The FDA has cleared autonomous AI devices before, starting with a tool that can detect diabetes-related eye problems in 2018 (the same tool received a CE mark in 2013). But autonomous radiology devices are more controversial. Professional organizations have spoken out against the idea: the American College of Radiology and the Radiological Society of North America published a joint letter in 2020 after an FDA workshop on artificial intelligence in medical imaging, saying that autonomous AI wasnt ready for clinical use. So far, they said, AI programs were too inconsistent and often didnt perform as well on groups of patients outside of the original environments they were built in.

Oxipit said in a statement that ChestLink made zero clinically relevant errors during pilot programs at multiple locations. When it is introduced into a new setting, the company said there should first be an audit of existing imaging programs. Then, the tool should be used under supervision for a period of time before it starts working autonomously.

The company said in a statement that it expects the first healthcare organizations to be using the autonomous tool by 2023.

View post:

First autonomous X-ray-analyzing AI is cleared in the EU - The Verge

Google has shared how its using artificial intelligence, including its restaurant-calling Duplex tech, to try and keep business hours up to date on Google Maps. The company says that if it is confident enough in the AIs prediction of what a businesss hours should be, it will update the information in Maps.

In a blog post, Google outlines the various factors its AI analyzes to determine whether it should do these updates. First, it looks at when the business profile was last updated, other similar shops hours, and Popular Times data to decide how likely it is that the hours are incorrect. For example: if Google sees that a lot of people visit the shop when its supposedly closed, that may be a red flag.

Googles post says that its AI looks at even more data if it determines the hours should be updated. Itll take into account information from the businesss website and can even scrape street view images (which may show business hours signs) to try and figure out when the business is open. Google says itll also check with actual humans, including Google Maps users and business owners, to verify the AIs predictions the company says it will even use Duplex in some countries to ask businesses about their hours directly.

Google spokesperson Genevieve Park told The Verge that Google will only publish business hours when we have a high degree of confidence that theyre accurate. If the AI thinks the hours may be incorrect but doesnt have a solid prediction, it adds a notice that the hours may have changed.

Park also said that Google doesnt explicitly tell users when hours were updated by its AI and explained that AI is used pretty much everywhere else in Google Maps. It seems like Googles pretty bullish on its AI-driven approach. In its post, the company says its on track to update the hours for over 20 million businesses around the globe in the next six months.

Google also says its piloting another use of AI in Maps to help keep speed limits up to date. In the US, itll try to see if its partners have taken images of stretches of road that have speed limit signs and will have AI help its operations team identify the sign and the speed limit posted on it.

While its no surprise that Googles using AI for these problems, it is interesting to see how many interlocking systems are involved. Theres computer vision, pattern recognition in location trends, and analyzing data about similar locations (which, of course, also involves figuring out what the similar locations even are), all to quietly try and keep up with how often businesses change their hours and make sure it knows the speed limit on certain stretches of road.

Here is the original post:

Google is using AI to update business hours that are out of date on Google Maps - The Verge

Chinese researchers last month unveiled an advanced AI system that enables low-cost commercial imaging satellites to become potent spy platforms.

In a paper published in the domestic peer-reviewed journal Fire Control and Command Control by Chinas state-owned defense industry, a Chinese team said its AI upgrade to its Jilin-1 satellite achieved a 95% precision rate in identifying small objects such as planes in the air or cars on the street, seven times more than previous technology.

The AI is reportedly capable of keeping track of moving objects, even if the object turns sharply or disappears into a tunnel.

According to Lin Cunbao of the Peoples Liberation Armys Space Engineering University, traditional satellite AI assumes it had made a mistake when it loses track of a target, achieving only a 14% success rate in analyzing satellite video.

In contrast, their new AI would estimate a moving targets direction based on experience and continue tracking it based on the most likely direction it would take. In addition, Lins paper mentioned that the AI could recapture the target as soon as it reappeared and added that their AI could work even better from space.

The Jilin-1 satellite, first launched in 2015, is Chinas first commercial Earth observation satellite. It is notably smaller than other spy satellites, weighing less than 100 kilograms.

China also showcased its AI-enhanced satellite reconnaissance capabilities in June last year when its Beijing-3 commercial satellite performed an in-depth scan of a 3,800 square kilometer area of San Francisco Bay in only 42 seconds at an altitude of 500 kilometers.

The images were sharp enough to allow military vehicles on the street to be identified, and what types of weapons they carried.

In contrast to traditional spy satellites that must remain stable while scanning an area of interest, the Beijing-3 rolled and yawed wildly, allowing it to scan far larger areas. The performance test over North America showed that Beijing-3 can take images with its body twisting up to 10 degrees per second, a capability not seen in previous satellites.

Beijing-3 is also claimed to have a response time 2-3 times faster than WorldView-4, the most advanced earth observation satellite developed by the United States. Also, Beijing-3s scanning band is 77% wider at 23 kilometers compared with WorldViews 13 kilometers, while weighing only half of its US counterpart.

Lead scientist Yang Fang stated in the peer-reviewed journal Spacecraft Engineering that China started relatively late on agile satellite technology, but achieved a large number of breakthroughs in a short period of time.

In 2020 Changguang Satellite, the manufacturer of the Jilin-1 Satellite, released satellite video footage of what appeared to be a fighter jet flying over a city, in an apparent showcase of the satellites tracking capabilities. However, it was not clear what type of fighter jet was being tracked in the video.

Future Chinese imaging satellites could be equipped with onboard AI and image processing capabilities, which would eliminate the need to broadcast data to ground stations for further analysis. This eliminates significant delays, especially if the satellites are tracking targets on the other side of the planet.

Chinese satellites will soon be able to stream live satellite footage to an end-users smartphone, a capability confined at present to the war rooms of leading military powers.

Chinas efforts to integrate military AI into its commercial satellites may be seen as an attempt to increase the survivability of its space-based intelligence, reconnaissance and surveillance (ISR) platforms via proliferation.

By 2025, China plans to launch the full constellation of 138 Jilin-1 satellites in orbit. Upgrading these satellites with onboard AI that considerably increases their imaging capabilities would increase their persistent monitoring capabilities and increase their survivability via proliferation against US and allied anti-satellite weapons, such as lasers, microwaves, electronic warfare, cyberattack and missiles through their sheer number.

The fact that these satellites were designed as civilian assets in the first place obfuscates the distinction between military and civilian assets in space.

Similar to how Chinas use of fishermen and civilian law enforcement agencies to assert its maritime claims in the South China Sea confounds the rules of engagement by other claimant states, the use of commercial satellites for potential military purposes may present similar problems in crafting the rules of engagement in outer space.

In addition, the dual-use nature of Chinas ISR satellite capabilities shows significant advancement in its military-civil fusion strategy, which aims to reorganize its science and technology enterprises to ensure that new innovations simultaneously advance economic and military development.

Continued here:

China's AI makes its satellites spies in the sky - Asia Times

As medicine continues to test automated machine learning tools, many hope that low-cost support tools will help narrow care gaps in countries with constrained resources. But new research suggests its those countries that are least represented in the data being used to design and test most clinical AI potentially making those gaps even wider.

Researchers have shown that AI tools often fail to perform when used in real-world hospitals. Its the problem of transferability: An algorithm trained on one patient population with a particular set of characteristics wont necessarily work well on another. Those failures have motivated a growing call for clinical AI to be both trained and validated on diverse patient data, with representation across spectrums of sex, age, race, ethnicity, and more.

But the patterns of global research investment mean that even if individual scientists make an effort to represent a range of patients, the field as a whole skews significantly toward just a few nationalities. In a review of more than 7,000 clinical AI papers, all published in 2019, researchers revealed more than half of the databases used in the work came from the U.S. and China, and high-income countries represented the majority of the remaining patient datasets.

advertisement

Look, we need to be much more diverse in terms of the datasets we use to create and validate these algorithms, said Leo Anthony Celi, first author of the paper in PLoS Digital Health (he is also the journals editor). The biggest concern now is that the algorithms that were building are only going to benefit the population thats contributing to the dataset. And none of that will have any value to those who carry the biggest burden of disease in this country, or in the world.

The skew in patient data isnt unexpected, given Chinese and American dominance in machine learning infrastructure and research. To create a dataset you need electronic health records, you need cloud storage, you need computer speed, computer power, said co-author William Mitchell, a clinical researcher and ophthalmology resident in Australia. So it makes sense that the U.S. and China are the ones that are in effect storing the most data. The survey also found Chinese and American researchers accounted for more than 40% of the clinical AI papers, as measured by the inferred nationality of first and last authors; its no surprise that researchers gravitate toward the patient data thats closest and easiest to access.

advertisement

But the risk posed by the global bias in patient representation makes it worth calling out and addressing those ingrained tendencies, the authors argue. Clinicians know that algorithms can perform differently in neighboring hospitals that serve different patient populations. They can even lose power over time within the same hospital, as subtle shifts in practice alter the data that flows into a tool. Between an institution from So Paulo and an institution in Boston, I think the differences are going to be much, much bigger, said Celi, who leads the Laboratory of Computational Physiology at MIT. Potentially, the scale and the magnitude of errors would be greater.

Clinician guidelines are already tailored to well-resourced countries, and a lack of diverse patient data only stands to widen global health care inequality. Most of the research that informs how we practice medicine is performed in a few rich countries, and then theres an assumption that whatever we learn from these studies and trials performed in a few rich countries will generalize to the rest of the world, said Celi. This is also going to be an issue if we dont change the trajectory with respect to the creation of artificial intelligence for health care.

The answer isnt straightforward, because nations that are resource-poor are also more likely to be data-poor. One popular research target for clinical AI in low-resourced settings is automated screening for eye disease. Using a portable fundus camera to image the eye, or even a smartphone camera, an algorithm could identify the signs of problems like diabetic retinopathy early enough to intervene. But as the authors note, 172 countries accounting for 3.5 billion people have no public ophthalmic data repository for researchers to draw from data deserts that frequently also affect other fields of medicine.

Thats why Celi and others are investing in programs to encourage data collection and pooling of machine learning resources in poorly-represented countries. One consortium is assembling multidisciplinary experts from Mexico, Chile, Argentina, and Brazil to identify best practices in data diplomacy, said Celi. It turns out the biggest challenge here is really the politics and economics of data, encouraging those with access to clinical data to open it up for local and international research rather than hoarding it for commercial purposes.

That work can also help double down on efforts to test existing models in areas with data disparities. If local data collection and curation isnt possible yet, validation can help ensure that algorithms trained in data-rich countries can, at least, be safely deployed in other settings. And along the way, those efforts can start to lay the groundwork for long-term data collection, and the ultimate growth of international data repositories.

By quantifying the international bias in AI research, Celi says, we just dont end up with things are pretty bad. The group hopes to use this as a baseline against which to measure improvement. Another recent paper led by Joe Zhang at Imperial College London detailed the creation of a dashboard that tracks the publication of clinical AI research, including the nationality of the first author on each paper. The first step to solving the problem is measuring it.

See original here:

More than half of data used in health care AI comes from the US, China - STAT

Why data and AI are the next step in education

Digital systems around the world generate a staggering 2.5 quintillion bytes of new data every day.1 While this information is generally stored in large data silos where it can be easily accessed by users, industries have been harvesting their data for years to make themselves more efficient and effective.

Artificial Intelligence (AI) enables data holders to transform a passive resource into a powerful catalyst for accelerated growth. For instance, when the state of Nebraska realized that it was spending approximately 655,000 staff hours per year to collect data reports from every school in the territoryan effort that yielded surprisingly few benefitsthe Nebraska Department of Education set out to build a Statewide Longitudinal Data System that would allow information to flow in near-real-time from hundreds of sources and deliver actionable insights to state, district, and school leaders; administrators; and educators.

Despite the noticeable positive impact, until recently, the education sector had been relatively slow to embrace digitalization and the use of data and AI to accelerate learning. However, COVID-19 created an urgent need for education systems to use their data to gain visibility into who was engaging in remote learning and where education was taking place. Education leaders the world over were motivated to take decisive action and schools began to make the transition to online learning as quickly as possible.

Throughout the COVID-19 response we understood clearly the importance of having data in order to measure the impact of this unprecedented disruption to education, observed Stefania Giannini, Assistant Director-General for Education at UNESCO.

Today, many institutions have settled into a blended or hybrid learning model and want to see what other benefits their digital framework can offer. Using cloud technology enables them to gain visibility and accelerate the impact of teaching and learning systems.

Education Insights, a feature in Microsoft Teams for Education, is a great example of this. It uses data analytics to keep educators informed of students engagement, learning progress, and well-being. A wide range of built-in digital apps and tools allows teachers to interact with learners on the platform and gain an overview of how well they are progressing, both at the class and individual levels.

One such tool is Reading Progress, a literacy solution that enables students to record themselves while reading aloud. The program makes a note of all the words that are challenging to the reader and provides visual aids and additional reading exercises to help them improve. Best of all, Reading Progress saves teachers hours of time spent evaluating students one at a time. It also allows teachers to take a personalized approach to teaching by addressing each students needs individually.

Given the power and centrality of literacy in conferring future outcomes, we are very proud of Reading Progress, and we are excited to continue to build on it and do even more, said Steve Liffick, Vice President Modern Life and Learning.

Among the key benefits of cloud technology are that it allows institutions to retain full ownership of their student data as well as receive expert support from partner technology companies on how to integrate security protocols and create governing policies around that data. Last year, UNESCOs member states adopted the first-ever global agreement on the ethics of AI. The document outlines a framework for the ethical use of AI including a chapter that is specifically focused on the ethics of AI in education.

UNESCO has been at the forefront of the international response to the global education crisis since the beginning, launching the Global Education Coalition in early 2020. The platform brought together more than 175 members from the UN family, civil society, academia, and the private sector to protect the right to education during the pandemic and beyond. Members are united under the coalitions three flagships: connectivity, teachers, and gender. Weve noticed that in many countries girls have been left behind, said Stefania. Filling the gender gap is something that UNESCO has been focusing on since the beginning of the pandemic.

In order to ensure that all children are able to receive the benefits of education analytics and AI, all children have to participate in digital learning observes Paige Johnson, Vice President, Microsoft Education Marketing. As long as some children are still operating in the analog world, you risk creating Big Data systems that leave those children out of the thinking and the work.

Another aspect to consider is that in order to benefit from big data systems, all students must take part in digital learningotherwise, education leaders run the risk of excluding certain learners from the data and the solutions such learning enables. This is why equipping every student with a digital device is the first step to implementing a successful data and AI strategy.

Helsinki was the first capital city to recognize the importance of having a digitalization strategy for education. In 2016, city officials set out to make Helsinki the most impactful place for learning in the world. Working with Microsoft, the citys Education Division used Azure to build a powerful AI hub capable of enhancing teaching and learning across a wide range of pedagogical use cases. The teams primary focus was to create a personalized learning experience for each of their students, all while improving learning outcomes and placing an emphasis on well-being.

Open Education Analytics is an open source program created by Microsoft to support every education systems unique journey with data and AI. The program was launched as a response to the urgent need for visibility into what was happening with education systems at the onset of the pandemic, especially as those systems moved more and more to digital learning platforms. We realized in that moment that we needed to accelerate our support for our customers data and AI journeys, said Maria Langworthy, Principal Program Manager at Microsoft.

The four components that make up the Open Education Analytics program are a set of open-source technical resources, a comprehensive curriculum on data engineering and data science training, Microsofts principles for responsible AI, and a global community of education systems developing shared use cases.

Each pillar is designed to solve specific challenges to digitalization and to empower education systems to navigate their way forward. As Maria noted, Its not just data and a lot of dashboards. Its about how you use this data to make better decisions, to better utilize your resources to really push learning progress.

Our goal with Microsoft Education Data and AI programs is to meet every education system where they are today and to help them move forward, [in order] to better leverage the data that they have using our modern data and AI services, said Louise Macquet, EMIS Cloud Business Development Lead, Microsoft MEA.

The Education Management Information System (EMIS) is designed to enable education systems to effectively collect, store, manage and report their data.

It does this through an open-source common data model for education that provides systematic consistency for data and supports education systems to develop applications and integrators more quickly to operate across multiple systems more easily. "The common data model was developed by Microsoft and founding partners to eliminate data silos for a connected engaged platform producing efficient and real-time data," explained Louise.

In order to empower education leaders across the world to accelerate the digital transformation of their systems and achieve meaningful impact in education, Microsoft created the Leaders in Digital Transformation of Education program. Join the program today and sign up for the Data and AI Accelerator to harness the power of your student learning data in real time for optimal results.

1https://techjury.net/blog/how-much-data-is-created-every-day/#gref

See the article here:

How data and AI are changing the world of education - Microsoft

SEGA is using advanced next-gen technologies to build its new Super Games including Unreal Engine 5, AI, and procedural FX tools like Houdini, company management has confirmed.

VIEW GALLERY - 2 IMAGES

SEGA has big plans for the future of gaming. The company created a new term called Super Games to describe its ambitious idea; Super Games are described as massive online-driven inter-connected titles that engage players on all platforms all over the world. Phantasy Star Online 2 is a great example of a Super Game.

SEGA wants to have multiple Super Games on the market by 2030 and is investing over $800 million into this new plan while also converging all of its game development segments to work on the multi-game initiative.

To make these games a reality, SEGA is utilizing new tech to develop its Super Games like Unreal Engine 5, which features advanced Nanite micro-polygon virtualized geometry and new Lumen lighting systems, alongside next-generation cloud-based development environments powered by Microsoft's Azure cloud network.

SEGA is also using Houdini, the same procedural generation FX tools that Epic Games used to build the city in its impressive The Matrix Awakens demo on PlayStation 5.

Here's what SEGA producer Masayoshi Kikuchi said about the Super Games development environment:

"We are also actively incorporating new technologies from the outside. In the fall of 2021, we announced that we would consider a business alliance with Microsoft, but our goal is to explore the possibilities of technology that SEGA cannot have and create new ones. To achieve this, we are positively promoting partnerships with various companies and intending to incorporate them into the game.

"As an example of design development, we are working on building a development flow that automatically generates objects using software called "Houdini". In modeling, we have also introduced technology such as creating CG from clothing patterns.

"We have prepared an environment where you can build a career in each work area, such as 3D models, motions, and effects. We will continue to actively invest in such environmental aspects."

"Development is basically done with Unreal Engine 5," SEGA general Katsuya Hisai said.

"In addition, in collaboration with start-up companies that have AI technology, we are also taking on challenges using AI, from the back end such as debugging to the front such as in-game camera work, live commentary, and automatic voice synthesis.

"Right now, I'm trying to figure out what can be achieved through trial and error on various technologies. From that point of view, I think there are many opportunities to come into contact with the latest technology."

SEGA hopes to have its first Super Game on the market by 2026.

Read more:

SEGA's new Super Games to use Unreal Engine 5, AI and Houdini - TweakTown

They can only dream of what its like to burst onto the field in The Big Show on Opening Day, but Purdue University outfielders Cam Thompson and Curtis Washington Jr. are among thousands of college baseball players with access to more data-juiced tech than ever to use in the hopes of getting to the majors. One of the tools their team has tested tracks and visualizes every joint in their bodies to measure and analyze their dynamic movements, helping them become a split-second faster on the base paths or gain an edge on runners when they throw home.

I was the slowest on the team, said Thompson in a video describing Purdues use of 3D Athlete Tracking, or 3DAT, technology developed by Intel, which captures video footage and applies computer vision and deep learning to digitize an individual players skeletal data and calculate biomechanics. The data and analytical insights gave Thompson and his coaches information revealing that he was bent over just slightly when launching himself from a base.

To the eye, you might not see this, but those first four or five steps were actually slowing him down, said John Madia, director of Baseball Player Development at Purdue.

After straightening up his running stance, Thompson said, It made me a whole lot faster, where Im close to the top in the fastest on the team now.

For college players like Thompson and Washington Jr. as well as pro athletes throughout sports, the use of data showing how their bodies move, breathe, sleep and recover from injury is becoming commonplace. In fact, while money was at the heart of the excruciatingly prolonged negotiations this winter between Major League Baseball and its players union, a clause in the final collective bargaining agreement addresses the data reflecting players bodies as another form of currency used to assess their value.

The new Collective Bargaining Agreement makes it illegal for the MLB and any of its teams to sell and/or license any players confidential medical information, personal biometric data or any non-public data.

3DAT technology captures video footage and applies computer vision and deep learning to digitize an individual players skeletal data and calculate biomechanics.Screenshots: Purdue University; Youtube

Right away, bettors and sportsbook companies saw the new data rules as a clear sign that players were recognizing the potential for their bodily data points to be used by gamblers to predict performance of specific players or teams. Some reports pointed out the fact that the MLB already inked a deal in 2018 that made MGM Resorts baseballs official gaming partner most including obligatory allusions to baseball betting scandals such as the 1919 Chicago Black Sox and former All-Star Pete Rose.

The use of data measuring players agility or injury recovery progress or revealing the impact of nutrition, sleep and hydration on their performance has implications not just for betting, but an athletes entire career trajectory. While wearables, apps and AI-based technologies can produce data and analysis that can literally change the game for some college players, they have little control and few rights over the data associated with the very bodies that dictate their futures.

Very few organizations cover it in their union contracts, said Kimberly Houser, a professor specializing in emerging technology law at the University of North Texas, who studies athlete biometric data use. And, because no law directly applies to the type of data generated by 3DAT or other devices used by athletes, she said, These device-makers arent really considering all of the results.

Even the MLBs recently settled contract with the MLB Players Association does not mention the device-makers facilitating data collection and analysis, Houser said. Without a separate contract between the league and the device-maker, she continued, They have this false sense that theyre being protected.

Intel has no official relationship with Major League Baseball. The company said it is exploring additional sports-related partnerships to bring its 3DAT technology to other players. For instance, athletic training company Exos piloted 3DAT with football players hoping to improve enough to be drafted to the NFL someday.

For Madia, access to the 3DAT system has nothing to do with betting. Its about helping the universitys players improve and giving its recruiters a leg up on elite schools, many of which not only can afford to outfit athletes with physical sensors to track their body movements, but already might collect and monitor their blood, urine, sweat or sleep patterns to evaluate their nutrition status in the hopes of maximizing performance and injury recovery time.

I look at dozens of things a day that I go, So what? How does this translate to winning? Madia said. From a recruiting standpoint, this is such a cool thing for Purdue to have.

In the past, old-school techniques such as timing a players speed using a watch or gauging the velocity of a throw with a radar gun only offered insights into which skills players needed to improve, but not how. It was a little bit of guessing at that point wheres my deficiency? said Madia. Its almost a diagnostic that hasnt been there to this degree, and I think were just scratching the surface on it.

Unlike earlier technologies that required players to wear sensors that could impede their natural movements, one benefit of Intels technology comes through its use of standard video footage. The system can apply its computer-vision AI even to videos captured using a mobile phone camera. And by tracking each persons full-body skeletal kinematics, it generates a slew of intricate data points.

For any recording, we can analyze 2,000-plus types of data including joint positions, velocities and accelerations, Breana Cappuccilli, an AI and sports technology product engineer at Intel working with 3DAT, told Protocol. She said the staff at Purdue, where she was a mechanical engineering doctoral candidate and graduate technical intern at Intel while working with 3DAT at the school, narrowed those down to around 10 data fields to devise metrics that could be easily communicated to players to help them improve base running.

When focusing on base running, we identified key metrics such as peak speed, angle of attack and stride length to help the players optimize their form, said Cappuccilli, who has since received her doctorate from Purdue.

The fact that 3DAT can capture a players knee movements while jumping and landing at a fast-enough frame rate using a mobile phone camera is key for AiScout, a company that incorporated the technology into its app used by amateur soccer players and professional clubs. Kids and older players shoot video footage with the app while performing standard speed and agility drills typically used by scouts to gauge a players skills.

Once Intel translates video footage into relevant joint-tracking points, AiScout which integrated 3DAT into its app in 2021 analyzes how those points moved to measure how well a player performs. Richard Felton-Thomas, director of Sport Science and chief operating officer at AiScout, said that although 3DAT captures image data more reliably than other systems hes tested, Intels technology could use some improvement when it comes to generating athlete data from video captured under poor lighting or weather conditions, or when a phone camera is much more than 10 meters away.

By giving players a tool to generate data showing how well they perform standard moves, the company aims to help expose them to scouts, whether or not they are part of the elite player networks where recruiters usually look for draftees. Normally someone goes out with a notepad and a pen, said Felton-Thomas. They go down the same networks year in and year out.

Indeed, its the data itself that could make or break a players career.

According to Felton-Thomas, the players themselves own the data generated using AiScout, and by choosing which teams virtual trials they enter through the app, they are the ones who decide which clubs or organizations can access their information. For example, Felton-Thomas said, We do not take the data of a player who has entered the Chelsea trials and then deliver that information to other clubs or organizations.

Players are beginning to recognize the value of the data reflecting their body movements and biomechanics, Felton-Thomas said. Its starting to be discussed more in particular, the onus of shifting the players data to being owned by that player or at least co-owned so they can use and monetize it themselves, he said.

Now, Intel wants other app-makers and tech providers outside the world of sports to integrate its 3DAT system for use in health care, physical therapy or fitness settings.

The uses for this technology go way beyond sports, said Jonathan Lee, director of Sports Performance Technology in Intels Olympic Technology Group, who said that in the past year the company has focused on bringing the technology to other app developers.

While Intel envisions a day when 3DAT can tell an everyday home-fitness buff whether her downward-dog pose or squat-thrust positioning needs adjustment, the company would not say whether any deals with non-sports-related partners are in the works.

Representatives of Intel and AiScout tout the potential benefits for AI-based phone apps to level the real-life and figurative playing fields for athletes and non-athletes alike. But future uses of the data gathered and created by these systems are largely unknown, leaving unanswered questions around data ownership, control and privacy risks.

Whether or how current laws apply to the type of data created in a system such as 3DAT is up to legal interpretation. For example, if a physician or team medical staff or another entity covered by the Health Insurance Portability and Accountability Act were using the data, it would likely be subject to protections associated with that federal law, said Kate Black, a partner in Hintze Laws health and biotech privacy group.

However, when makers of wearables or other devices collect and analyze data about someones body or health characteristics, the data might not be covered by HIPAA, said Houser.

If someone can infer that an athletes physicality is degrading, they could use that against an athlete.

Its also unclear how state privacy or biometric data laws might apply. For example, while the Biometric Information Privacy Act in Illinois is centered on the use of identifiable physical characteristics such as retina/iris scans, voiceprints and fingerprints, it also covers a scan of hand or face geometry.

Its very likely in my opinion that cases will be brought under the state biometric laws, Black said. For instance, she said data reflecting the way a baseball player holds a bat or grips a ball could be deemed relevant according to the Illinois laws hand-scan language, possibly inspiring people to sue for damages under its private right of action.

Both Black and Houser said player data gathered, combined and analyzed over time could create unintended consequences for athletes.

Data silos are continuing to break down, Black said. Putting together health or fitness assessments of an individual that combines their medical history, X-rays, biometric data, genetic information I dont think its too far off from creating a performance score or an individual risk score that could be used to inform [an athletes] recruitment to be used in any sport.

Intels mission for its 3DAT technology is to improve athletes performance and rehabilitation, said a company spokesperson. The use of this data is only intended for improving how people move more effectively, efficiently and safely. Intel and its partners abide by strict data privacy policies to ensure data is protected end-to-end. The spokesperson noted that Intel has no intention of sharing data derived from 3DAT with other entities for sports betting or to monetize it in any way. Intels customers own the data generated by 3DAT, he said.

But for partners such as Purdue, the very definition of the data that 3DAT generates is muddled.

At this point we/Purdue are not limited [in] our ability to utilize and/or share data from the 3DAT technology on our athletes, wrote Madia in an email. He made a distinction between health data such as athletes' dietary information or muscle-to-fat ratio data protected under HIPAA and the information generated through 3DAT.

The 3DAT biomechanical [data] which demonstrates posture, explosiveness, technique, is ours with no restrictions, he said.

There are several risks associated with unwanted or unrestricted use or access to biometric data from denial of health coverage based on information indicating an illness or injury to use by law enforcement or immigration authorities. But there are particular risks to athletes.

The concern in this is that parties could use this information in contract negotiations, trade decisions, and other ways that the athlete was not expecting and may not desire, wrote Houser and John Holden, an associate professor focused on sports law at Oklahoma State University, in a 2021 research paper investigating the subject of athletic biometric data.

If someone can infer that an athletes physicality is degrading, they could use that against an athlete, Houser told Protocol. And, because there is often no bargaining organization representing college players in such situations, they could be compelled to waive their data rights in exchange for maintaining their scholarships, she continued. Theyre in a much worse position than professional athletes.

Read more from the original source:

AI could help college baseball players reach the majors, but with little control over their biometric data - Protocol

Moritz Stefaner has long been obsessed with food. As a designer, he has even used meals to visualize data about everything from ethnic diversity to scientific funding. So when he encountered an AI that generates realistic pictures from words, he went hog wild.

[Image: Midjourney]Stefaner typed words like Michelin star chef, deconstructed, and amuse-gueule into the generator, hoping to evoke the intricate plating of fine dining establishments like Eleven Madison Park or Alinea. And, suffice to say, his plan worked. The images he created are completely convincing plates that you could imagine being served at any 3-Michelin-starred restaurant. That is, until you look a bit closer and you realize that the individual components on the dish often dont even exist in real life.

[Image: Midjourney]Squinting, I see the bones of a soft-boiled egg, seaweed, microgreens, sauces, and gels. In one frame, I swear I see the same candied moss I actually ate at the Chicago fine dining establishment Elizabeth. In another, I see a chocolate dessert sitting on a mound of coffee groundsa stones throw from a dish I ate at Dominique Ansel Kitchen. But for the most part, its superbly convincing fictionand what happens when an AIs style vastly outpaces its substance.

[Image: Midjourney]The most surprising thing to me was how well the system deals with really poetic descriptions, writes Stefaner via email. It goes way beyond just capturing objects in certain styles, towards capturing a whole vibe.

[Image: Midjourney]The AI was built by Midjourney, a self-ascribed research lab focused on expanding the imaginative powers of the human species. Much like GLIDE, Disco Diffusion, and Dall-E, Midjourneys AI model generates images from words. But Midjourney makes the process particularly simple. You dont have to understand code or set up anything special to use this AI. Instead, it is hosted on a private Discord, so creating an image is literally as easy as typing it.

[Image: Midjourney]The input to these networks are short texts called prompts. The art of the prompt is really becoming a key skill in interacting with these models. Similar to learning how to nudge a search engine to surface the right results, the prompt artist learns to use the right combination of words to achieve the desired effects, Stefaner says. For instance, one can add drawn by Picasso or in the style of Keith Haring to evoke style modifiers that mimic an artists style.

Stefaner focused on words like fine dining, and stylistically he tagged many of his prompts with dof, which stands for depth of field, and refers to the partially in-focus images that appear to be shot by a traditional camera and are a hallmark of fine dining photography.

No doubt, the way these images are framedthe angles at which the camera seems to be taking themhelps them seem convincing. Thats key because most of these foods are at least a little alien; they are almost-foods, if you will. Pasta dishes look more like thinly sliced banana peels. Fish looks like salmon, if its skin were marbled into its flesh. These oddities arent always as gross as they might sound. One anonymous fine dining plate looks like a cross between prawn shells and flowers. Its downright beautiful, and just the sort of meticulous surprise you hope to encounter when dropping hundreds of dollars on a tasting menu.

I wish I could say the same about what appears to be sea scallops. Seared on top, they melt like a Salvador Dal painting into the plate. Are they made of ice cream? Might they be a foam? My brain tries to make sense of it all until I remember, theres no sense to be made. Im looking at an AI hallucination.

[Image: Midjourney]Its almost like an alien life form observed us and tried to imitate and blend in the best it could, without really understanding what is going on, Stefaner says. This strangely familiar unfamiliar feeling is a bit unsettling, but can also really trigger creativity. We are pattern-seeking animals, always searching for meaning, so we really try to figure out what these dishes could be, what they could taste like, even though they dont quite make sense to us.

[Image: Midjourney]These hallucinations, of course, are trained into the software, which was fed countless labeled images to understand how to draw the objects. And there is no better window into the AIs superficial logic than in Mortizs fine dining high end Michelin star closeup burgeron top, its a pile of rare ground beef, capped with a shiny brioche bun. But on the bottom, where the other bun should be? Thats a coral-like pile of something vaguely edible. In other words, the burger starts at Red Robin and ends with Noma. And while its funny, this image also demonstrates how little these AI models understand about the content they generate.

[Image: Midjourney]In any case, Stefaners images are captivating to behold. They also push us to ask, Whats next? Thus far, weve seen art imitate life. But next, we might see life imitate art.

Id love to do creative sessions with ambitious chefs to generate inspiring images, based on new prompts (or their existing menus!) and then see if we can together reverse-engineer them into successful dishes, Stefaner says. Its a new type of agent you can inject in your design process to generate completely new, oblique ideas.

Read this article:

These 3-Michelin-starred plates were invented by AI. The food doesn't even exist - Fast Company

Lilt, a provider of AI-powered business translation software, today announced that it raised $55 million in a Series C round led by Four Rivers, joined by new investors Sorenson Capital, CLEAR Ventures and Wipro Ventures. The company says that it plans to use the capital to expand its R&D efforts as well as its customer footprint and engineering teams.

Lilt [aims to] build a solution that [will] combine the best of human ingenuity with machine efficiency, CEO Spence Green told TechCrunch via email. This new funding will [reduce our] unit economics [to make] translation more affordable for all businesses. It will also [enable us to add] a sales team to our existing production team in Asia. We are in three regions the U.S., Europe, the Middle East and Africa (EMEA) and Asia and look to have both sales and production teams in each of these regions.

San Francisco, Calfornia-based Lilt was co-founded by Green and John DeNero in 2015. Green is a former Northrop Grumman software engineer who later worked as a research intern on the Google Translate team, developing an AI language system for improving English-to-Arabic translations. DeNero was previously a senior research scientist at Google, mostly on the Google Translate side, and a teaching professor at the University of California, Berkeley.

15 years ago, I was living in the Middle East, where you make less money if you speak anything other than English. I have never been exposed to that kind of disparity before and the disadvantage was extremely frustrating, Green told TechCrunch. I then returned to the States, went to grad school and started working on Google Translate, where I met [DeNero]. Our mission is to make the worlds information accessible to everyone regardless of where they were born or which language they speak.

To translate marketing, support and e-commerce documents and webpages Lilts principal workload Lilt uses a combination of human translators and tools including hotkeys, style guides and an AI translation engine. Green says that the platform supports around 40 languages and offers custom term bases and lexicons, which show translators a range of possible translations for a given word.

The aforementioned AI engine, meanwhile which is regularly trained on fresh data, including feedback from Lilts translators analyzes translation data to make recommendations. But the translators have the final say.

AI and machine learning are helping automating the process around enterprise translation, but you cant automate it all thats why we are a human-in-the-loop process, Green said. We are leaving creative and emotional elements of translation to humans while automating the tedious and repetitive elements. This helps with the unit economics of our business and allows for businesses to apply translation across all customer touch points.

Using Lilts platform, customers can assign translators and reviewers, track due dates and keep tabs on ongoing translation job progress. After signing an annual contract with Lilt, customers can use the services API and connectors to funnel text for translation from platforms including Slack, WordPress, GitHub, Salesforce, Zendesk and Adobe Marketing Cloud.

Translators are paid a fixed hourly rate, negotiated individually. They must earn at least $20 for rendering linguistic services through the platform to cash out, which can include review as well as translation. Lilt tracks hours automatically, counting only time spent actively translating and reviewing content and not time spent conducting external research beyond Lilts standard work limits (30 seconds for translation per segment and 50 seconds for review per segment).

According to a recent Salesforce survey, the average consumer now uses ten channels including social media and SMS to communicate with businesses. Yet the average company supports relatively few languages, with a July 2020 study from Stripe finding that 74% of European e-commerce websites hadnt translated their checkout pages into local languages.

This is where Green sees opportunity, despite competition from rivals like Unbabel. Grand View Research anticipatesthe machine translation market will be worth $983.3 million by 2022.

Already, Lilt which has a workforce of over 150 people claims to have customers in Intel, Emerson, Juniper Networks and Orca Security and others across the education, crypto, technology, defense and intelligence sectors. Its only becoming more vital to ensure an end-to-end multilingual customer experience, Green added.

Existing investor Sequoia Capital, Intel Capital, Redpoint Ventures and XSeed Capital also participated in Lilts Series C round. It brings the startups total capital raised to $92.5 million.

More here:

Lilt raises $55M to bolster its business-focused AI translation platform - TechCrunch