Archive for the ‘Machine Learning’ Category

Ames Lab, Texas A&M team develop AI tool for discovery and prediction of new rare-earth compounds – Green Car Congress

Researchers from Ames Laboratory and Texas A&M University have trained a machine-learning (ML) model to assess the stability of new rare-earth compounds. The framework they developed builds on current state-of-the-art methods for experimenting with compounds and understanding chemical instabilities. A paper on their work is published in Acta Materialia.

Machine learning is really important here because when we are talking about new compositions, ordered materials are all very well known to everyone in the rare earth community. However, when you add disorder to known materials, its very different. The number of compositions becomes significantly larger, often thousands or millions, and you cannot investigate all the possible combinations using theory or experiments.

Ames Laboratory Scientist Prashant Singh, corresponding author

The approach is based on machine learning (ML), a form of artificial intelligence (AI), which is driven by computer algorithms that improve through data usage and experience. Researchers used the upgraded Ames Laboratory Rare Earth database (RIC 2.0) and high-throughput density-functional theory (DFT) to build the foundation for their ML model.

High-throughput screening is a computational scheme that allows a researcher to test hundreds of models quickly. DFT is a quantum mechanical method used to investigate thermodynamic and electronic properties of many body systems. Based on this collection of information, the developed ML model uses regression learning to assess phase stability of compounds.

Singh explained that the material analysis is based on a discrete feedback loop in which the AI/ML model is updated using new DFT database based on real-time structural and phase information obtained from experiments. This process ensures that information is carried from one step to the next and reduces the chance of making mistakes.

Singh et al.

Yaroslav Mudryk, the project supervisor, said that the framework was designed to explore rare earth compounds because of their technological importance, but its application is not limited to rare-earths research. The same approach can be used to train an ML model to predict magnetic properties of compounds, process controls for transformative manufacturing, and optimize mechanical behaviors.

Its not really meant to discover a particular compound. It was, how do we design a new approach or a new tool for discovery and prediction of rare earth compounds? And thats what we did.

Yaroslav Mudryk

Mudryk emphasized that this work is just the beginning. The team is exploring the full potential of this method, but they are optimistic that there will be a wide range of applications for the framework in the future.

This work was supported by Laboratory Directed Research and Development Program (LDRD) program at Ames Laboratory.

Resources

Prashant Singh, Tyler Del Rose, Guillermo Vazquez, Raymundo Arroyave, Yaroslav Mudryk (2022) Machine-learning enabled thermodynamic model for the design of new rare-earth compounds, Acta Materialia, Volume 229,117759 doi: 10.1016/j.actamat.2022.117759

The rest is here:
Ames Lab, Texas A&M team develop AI tool for discovery and prediction of new rare-earth compounds - Green Car Congress

Machine Learning Chip Market: Latest Trends and Forecast Analysis Up to 2029 The Sabre – The Sabre

This market report comprises of the most recent market information with which companies can have in depth analysis of industry and future trends. By applying market intelligence for this business report, industry experts assess strategic options, outline successful action plans and support companies with critical bottom-line decisions. Competitive analysis studies of this credible report helps to get ideas about the strategies of key players in the market. Not to mention, the scope of This market research report can be broadened from market scenarios to comparative pricing between major players, cost and profit of the specified market regions.

Get the Sample of this Report with Detail TOC and List ofFigures@https://www.databridgemarketresearch.com/request-a-sample/?dbmr=global-machine-learning-chip-market

Machine Learning Chip Market is expected to reach USD 72.45 billion by 2027 witnessing market growth with the rate of 40.60% in the forecast period of 2020 to 2027.

Introduction of quantum computing, rising applications of machine learning in various industries, adoption of artificial intelligence across the globe, are some of the factors that will likely to enhance the growth of the machine learning chip market in the forecast period of 2020-2027. On the other hand, growing smart cities and smart homes, adoption of internet of things worldwide, technological advancement which will further boost various opportunities that will lead to the growth of the machine learning chip market in the above mentioned forecast period.

Lack of skilled workforce along with phobia related to artificial intelligence are acting as market restraints for machine learning chip in the above mentioned forecaster period.

We provide a detailed analysis of key players operating in the Machine Learning Chip Market:

North America will dominate the machine learning chip market due to the prevalence of majority of manufacturers while Europe will expect to grow in the forecast period of 2020-2027 due to the adoption of advanced technology.

Market Segments Covered:

By Chip Type

Technology

Industry Vertical

Machine Learning Chip Market Country Level Analysis

Machine learning chip market is analysed and market size, volume information is provided by country, chip type, technology and industry vertical as referenced above.

The countries covered in the machine learning chip market report are U.S., Canada and Mexico in North America, Brazil, Argentina and Rest of South America as part of South America, Germany, Italy, U.K., France, Spain, Netherlands, Belgium, Switzerland, Turkey, Russia, Rest of Europe in Europe, Japan, China, India, South Korea, Australia, Singapore, Malaysia, Thailand, Indonesia, Philippines, Rest of Asia-Pacific (APAC) in Asia-Pacific (APAC), Saudi Arabia, U.A.E, South Africa, Egypt, Israel, Rest of Middle East and Africa (MEA) as a part of Middle East and Africa (MEA).

To get Incredible Discounts on this Premium Report, Click Here @https://www.databridgemarketresearch.com/checkout/buy/enterprise/global-machine-learning-chip-market

Rapid Business Growth Factors

In addition, the market is growing at a fast pace and the report shows us that there are a couple of key factors behind that. The most important factor thats helping the market grow faster than usual is the tough competition.

Competitive Landscape and Machine Learning Chip Market Share Analysis

Machine learning chip market competitive landscape provides details by competitor. Details included are company overview, company financials, revenue generated, market potential, investment in research and development, new market initiatives, regional presence, company strengths and weaknesses, product launch, product width and breadth, application dominance. The above data points provided are only related to the companies focus related to machine learning chip market.

Table of Content:

Part 01: Executive Summary

Part 02: Scope of the Report

Part 03: Research Methodology

Part 04: Machine Learning Chip Market Landscape

Part 05: Market Sizing

More.TOC.. ..Continue

Based on geography, the global Machine Learning Chip market report covers data points for 28 countries across multiple geographies namely

Browse TOC with selected illustrations and example pages of Global Machine Learning Chip Market @https://www.databridgemarketresearch.com/toc/?dbmr=global-machine-learning-chip-market

Key questions answered in this report

Get in-depth details about factors influencing the market shares of the Americas, APAC, and EMEA?

Top Trending Reports:

About Data Bridge Market Research:

Data Bridge Market Research set forth itself as an unconventional and neoteric Market research and consulting firm with unparalleled level of resilience and integrated approaches. We are determined to unearth the best market opportunities and foster efficient information for your business to thrive in the market. Data Bridge endeavors to provide appropriate solutions to the complex business challenges and initiates an effortless decision-making process.

Contact:

Data Bridge Market Research

US: +1 888 387 2818

UK: +44 208 089 1725

Hong Kong: +852 8192 7475

Corporatesales@databridgemarketresearch.com

Read this article:
Machine Learning Chip Market: Latest Trends and Forecast Analysis Up to 2029 The Sabre - The Sabre

Machine Learning Reimagines the Building Blocks of Computing – Quanta Magazine

Algorithms the chunks of code that allow programs to sort, filter and combine data, among other things are the standard tools of modern computing. Like tiny gears inside a watch, algorithms execute well-defined tasks within more complicated programs.

Theyre ubiquitous, and in part because of this, theyve been painstakingly optimized over time. When a programmer needs to sort a list, for example, theyll reach for a standard sort algorithm thats been used for decades.

Now researchers are taking a fresh look at traditional algorithms, using the branch of artificial intelligence known as machine learning. Their approach, called algorithms with predictions, takes advantage of the insights machine learning tools can provide into the data that traditional algorithms handle. These tools have, in a real way, rejuvenated research into basic algorithms.

Machine learning and traditional algorithms are two substantially different ways of computing, and algorithms with predictions is a way to bridge the two, said Piotr Indyk, a computer scientist at the Massachusetts Institute of Technology. Its a way to combine these two quite different threads.

The recent explosion of interest in this approach began in 2018 with a paper by Tim Kraska, a computer scientist at MIT, and a team of Google researchers. In it, the authors suggested that machine learning could improve a well-studied traditional algorithm called a Bloom filter, which solves a straightforward but daunting problem.

Imagine you run your companys IT department and you need to check if your employees are going to websites that pose a security risk. Naively, you might think youll need to check every site they visit against a blacklist of known sites. If the list is huge (as is likely the case for undesirable sites on the internet), the problem becomes unwieldly you cant check every site against a huge list in the tiny amount of time before a webpage loads.

The Bloom filter provides a solution, allowing you to quickly and accurately check whether any particular sites address, or URL, is on the blacklist. It does this by essentially compressing the huge list into a smaller list that offers some specific guarantees.

Bloom filters never produce false negatives if they say the site is bad, its bad. However, they can produce false positives, so perhaps your employees wont be able to visit some sites they should have access to. Thats because they trade some accuracy for an enormous amount of data compression a trick called lossy compression. The more that Bloom filters compress the original data, the less accurate they are, but the more space they save.

To a simple Bloom filter, every website is equally suspicious until its confirmed to not be on the list. But not all websites are created equal: Some are more likely than others to wind up on a blacklist, simply because of details like their domain or the words in their URL. People understand this intuitively, which is why you likely read URLs to make sure theyre safe before you click on them.

Kraskas team developed an algorithm that can also apply this kind of logic. They called it a learned Bloom filter, and it combines a small Bloom filter with a recurrent neural network (RNN) a machine learning model that learns what malicious URLs look like after being exposed to hundreds of thousands of safe and unsafe websites.

When the learned Bloom filter checks a website, the RNN acts first and uses its training to determine if the site is on the blacklist. If the RNN says its on the list, the learned Bloom filter rejects it. But if the RNN says the site isnt on the list, then the small Bloom filter gets a turn, accurately but unthinkingly searching its compressed websites.

By putting the Bloom filter at the end of the process and giving it the final say, the researchers made sure that learned Bloom filters can still guarantee no false negatives. But because the RNN pre-filters true positives using what its learned, the small Bloom filter acts more as a backup, keeping its false positives to a minimum as well. A benign website that could have been blocked by a larger Bloom filter can now get past the more accurate learned Bloom filter. Effectively, Kraska and his team found a way to take advantage of two proven but traditionally separate ways of approaching the same problem to achieve faster, more accurate results.

Kraskas team showed that the new approach worked, but they didnt formalize why. That task fell to Michael Mitzenmacher, an expert on Bloom filters at Harvard University, who found Kraskas paper innovative and exciting, but also fundamentally unsatisfying. They run experiments saying their algorithms work better. But what exactly does that mean? he asked. How do we know?

In 2019, Mitzenmacher put forward a formal definition of a learned Bloom filter and analyzed its mathematical properties, providing a theory that explained exactly how it worked. And whereas Kraska and his team showed that it could work in one case, Mitzenmacher proved it could always work.

Mitzenmacher also improved the learned Bloom filters. He showed that adding another standard Bloom filter to the process, this time before the RNN, can pre-filter negative cases and make the classifiers job easier. He then proved it was an improvement using the theory he developed.

The early days of algorithms with predictions have proceeded along this cyclical track innovative ideas, like the learned Bloom filters, inspire rigorous mathematical results and understanding, which in turn lead to more new ideas. In the past few years, researchers have shown how to incorporate algorithms with predictions into scheduling algorithms, chip design and DNA-sequence searches.

In addition to performance gains, the field also advances an approach to computer science thats growing in popularity: making algorithms more efficient by designing them for typical uses.

Currently, computer scientists often design their algorithms to succeed under the most difficult scenario one designed by an adversary trying to stump them. For example, imagine trying to check the safety of a website about computer viruses. The website may be benign, but it includes computer virus in the URL and page title. Its confusing enough to trip up even sophisticated algorithms.

Indyk calls this a paranoid approach. In real life, he said, inputs are not generally generated by adversaries. Most of the websites employees visit, for example, arent as tricky as our hypothetical virus page, so theyll be easier for an algorithm to classify. By ignoring the worst-case scenarios, researchers can design algorithms tailored to the situations theyll likely encounter. For example, while databases currently treat all data equally, algorithms with predictions could lead to databases that structure their data storage based on their contents and uses.

And this is still only the beginning, as programs that use machine learning to augment their algorithms typically only do so in a limited way. Like the learned Bloom filter, most of these new structures only incorporate a single machine learning element. Kraska imagines an entire system built up from several separate pieces, each of which relies on algorithms with predictions and whose interactions are regulated by prediction-enhanced components.

Taking advantage of that will impact a lot of different areas, Kraska said.

Here is the original post:
Machine Learning Reimagines the Building Blocks of Computing - Quanta Magazine

Military researchers to apply artificial intelligence (AI) and machine learning to combat medical triage – Military & Aerospace Electronics

ARLINGTON, Va. U.S. military researchers are asking industry to develop artificial intelligence (AI) and machine learning technologies for difficult jobs like combat medical triage, which refers to sorting wounded warfighters according to their need for medical attention.

Officials of the U.S. Defense Advanced Research Projects Agency (DARPA) in Arlington, Va., issued a broad agency announcement (HR001122S0031) this week for the In the Moment (ITM) project.

DARPA researchers are asking industry to develop algorithmic decision-makers that can help humans with decision-making in difficult domains like combat medical triage.

Difficult domains are where trusted decision-makers disagree; no right answer exists; and uncertainty, time-pressure, resource limitations, and conflicting values create significant decision-making challenges. Other examples include first response and disaster relief.

Related: Top technology challenges this decade for the warfighter

The DARPA ITM project focuses on two areas: small unit triage in austere environments, and mass casualty triage. ITM seeks to develop techniques that enable building, evaluating, and fielding trusted algorithmic decision-makers for mission-critical operations where there is no right answer and, consequently, ground truth does not exist.

Researchers are looking for capabilities that:

-- quantify algorithmic decision-makers with key decision-making attributes of trusted humans;

-- incorporate key human decision-maker attributes into more human-aligned, trusted algorithms;

-- enable the evaluation of human-aligned algorithms in difficult domains where humans disagree and there is no right outcome; and

Difficult decisions occur when the decision-maker is confronted with challenges that include too many or too few options, too much or too little information, uncertainty about the consequences of decisions, and uncertainty about the value of foreseeable outcomes.

ITM seeks to develop AI and machine learning algorithms based on key human attributes as the basis for trust in algorithmic decision-makers, as well as a computational framework for key human attributes and an alignment score match the algorithmic decision-maker to key human decision-makers.

Related: Simulation and mission rehearsal relies on state-of-the-art computing

ITM is interested in the notion of trust, or the willingness of a human to delegate difficult decision-making to AI computers. The project also will focus on human-off-the-loop, algorithmic decision-making in difficult domains to understand the limits of such a computational framework.

ITM is 3.5-year, two-phase program that focuses on four technical areas: decision-maker characterization; human-aligned algorithms; evaluation; and policy and practice.

Decision-maker characterization seeks to develop technologies that identify and model key decision-making attributes of trusted humans to produce a quantitative decision-maker alignment score.

Human-aligned algorithms should be able to balance situational information with a preference for key decision-maker attributes. Evaluation will assess the willingness of humans to delegate difficult decisions to AI computers.

Related: The next 'new frontier' of artificial intelligence

Policy and practice will develop recommendations for how military leaders can update policies to take advantage of AI and machine learning in combat medical triage.

Companies interested should upload abstracts by 30 March 2022, and proposals by 17 May 2022 to the DARPA BAA website at https://baa.darpa.mil/.

Email questions or concerns to Matt Turek, the DARPA ITM program manager, at ITM@darpa.mil. More information is online at https://sam.gov/opp/baae2217401748dbaeb89a08044d6998/view.

More here:
Military researchers to apply artificial intelligence (AI) and machine learning to combat medical triage - Military & Aerospace Electronics

Machine learning, AI can help ease the trend of physician burnout – Healthcare Finance News

Dr. Steven Waldren, vice president and chief informatics officer at the American Academy of Family Physicians, right, and Dr. Kamel Sadek, director of informatics at Village Medical, speak at the HIMSS22 conference in Orlando.

Photo: Jeff Lagasse/Healthcare Finance News

ORLANDO, Fla. Even before COVID-19 made the business of healthcare a nightmare for countless physicians and clinicians, burnout was a prevalent issue. And even the slow, still-ongoing emergence into normalcy hasn't been enough to ease this trend: Clerical burdens, including clinical documentation, are a major contributor.

But for primary care physicians in particular, a new class of technology, including AI-powered digital assistants, is improving their capacity and capability, while reducing their administrative and cognitive burden.

Dr. Steven Waldren, vice president and chief informatics officer at the American Academy of Family Physicians, cited data showing that the average patient visit to a PCP takes about 18 minutes, and of that time, 27% is dedicated to face-to-face time with a patient. Forty-nine percent is consumed by EHR and desk work.

"A lot of the technology is not really designed to connect with patients," said Waldren. "Physicians and nurses don't necessarily give EHRs a passing grade."

The three main challenges to primary care, he said, are clerical burden, value-based payments and AI and machine learning, which fundamentally change what it means to be a physician. But innovations in the marketplace can be applied to the clinical setting, said Waldren, and if they focus on initial contact, comprehensiveness, and coordinated and continuous care, practices can improve both costs and quality while tackling the burnout issue.

In 2018, AAFP sought to do just that, developing processes to assess the value of AI-based solutions for primary care. The group called them "innovation labs," with the goal of gauging the value of emerging technology solutions.

"We have a lot of innovators in family medicine and primary care who are doing great things," said Waldren. "We try to help our doctors find a doctor like them and a practice like theirs. They have to find peers who have been there, done that and been successful. If you don't do that you're going to have a hard time doing this."

The innovation labs identify products that address the challenges of EHRs and clerical burden, and they're geared to be adaptable and usable by family physicians.

To reduce EHR burden, the labs pinpointed a voice-assisted AI assistant for documentation. Instead of the typical voice recognition technology that simply transcribes a doctor's dictations onto a screen, the program understands commands and can respond to them, removing dictation from the equation entirely.

In testing this particular piece of technology, the labs gathered 132 members from small practices across 40 states, with all reporting much higher satisfaction. That's largely because of a 50% reduction in documentation time.

"It's geared toward those with documentation burden," said Waldren. "Its key features arte integration into the EHR, mobility and it's affordable and adaptable at about $200 per month."

For the problem of clerical burden, the labs identified technology that resulted in a 70% decrease in physician prep time, and a 38% increase in RAF scores. Physicians ranked it 9.6 out of 10.

One of those physicians, Dr. Kamel Sadek, director of informatics at Village Medical, joined Waldren on the stage and testified to the solution's efficacy.

"It's been a year since we tested the product," said Sadek. "We just adopted it, and within 10 days, we had 230 physicians who wanted to start using the product. We're on the onboarding process as we speak. Many physicians want to use the dictation. So far it's been a very, very successful pilot study and adoption.

"You're not connected to the computers," he said. "You can use your phone, iPad, laptop, youcan use it from your car, or while your kid is playing soccer. It's bidirectional. That gives you the brief information of what you've done, and based on that you can finish the note."

What convinced Sadek to utilize this approach was the efficacy and the practical aspect of being able to dictate from anywhere, to say nothing of reducing wasted time.

"With all of these things we have to right now in the notes, with value-based care, you have to monitor a lot of things that come from different departments," said Sadek. "In order to perform well, you have to look at all these things."

According to Waldren, health IT hasn't been very consumer-friendly to date. But newer apps are being built by companies that have experience in the more user-friendly consumer space, so the situation is beginning to improve. The best part is that they don't require a large up-front investment or complicated integration.

"One of the biggest challenges of getting a solution in is our doctors are so burned out," said Waldren, "and they don't want to look at something and decide if this is something that's going to work."

Twitter:@JELagasseEmail the writer:jeff.lagasse@himssmedia.com

View original post here:
Machine learning, AI can help ease the trend of physician burnout - Healthcare Finance News