Category Archives: Ai

The Mulholland family at home: Kyle, Conor, Melissa and Emma. (Photo by Scott Eklund/Red Box Pictures)

Conor, now 5, is already benefitting from AI in his life. After undergoing 12 surgeries by the age of 2 to deal with the effects caused by PUV and related issues, he was diagnosed with autism at age 3, and has had difficulty speaking. He is working with speech therapists, and is also benefitting from an app called Helpicto that uses AI to convert spoken words into a series of images.

Created by French company Equadex, the app, too, came from the heart several of the companys employees have personal experiences with autism. Equadex created Helpicto using Azure Cognitive Services and the Microsoft Azure cloud platform.

Its very hard to have a child not be able to speak or communicate with you, Mulholland says. He was 4 when he first said, Mama. When he did, it was amazing to hear it.

Conor also learned how to say Dada and Amma, or Emma, for his 7-year-old sister, who dotes on him.

Its really sweet to see how caring she is and how much she wants him to be successful, Mulholland says. She helps him practice on words. I can see her being in a field of study someday thats very focused on helping others, whether its as a teacher or doctor, or something along those lines. Her life will be forever changed because of having a brother like him.

Mulholland says she is humbled by all the support she has had from her husband, Kyle, an accountant who stays at home with Conor, to Microsoft for giving her a platform to tell her story, to the companies that want to hear and embrace it.

I always encourage people, Dont pigeonhole yourself, think of ways that you can really harness technology to drive greater good, because sometimes those solutions are right in front of you, she says. And imagine how great of a world we could live in if we had more stories like this.

Top photo: Melissa Mulholland with son Conor. (Photo by Scott Eklund/Red Box Pictures)

View post:

How one Microsoft mom inspired health care companies to embrace the life-saving potential of AI | Transform - Microsoft

Monitors display a video showing facial recognition software in use at the headquarters of the artificial intelligence company Megvii, in Beijing, May 10, 2018. Beijing is putting billions of dollars behind facial recognition and other technologies to track and control its citizens.

Gilles Sabri | The New York Times

The world is facing its biggest health crisis in decades but one of the world's most promising technologies artificial intelligence (AI) isn't playing the major role some may have hoped for.

Renowned AI labs at the likes of DeepMind, OpenAI, Facebook AI Research, and Microsoft have remained relatively quiet as the coronavirus has spread around the world.

"It's fascinating how quiet it is," said Neil Lawrence, the former director of machine learning at Amazon Cambridge.

"This (pandemic) is showing what bulls--t most AI hype is. It's great and it will be useful one day but it's not surprising in a pandemic that we fall back on tried and tested techniques."

Those techniques include good, old-fashioned statistical techniques and mathematical models. The latter is used to create epidemiological models, which predict how a disease will spread through a population. Right now, these are far more useful than fields of AI like reinforcement learning and natural-language processing.

Of course, there are a few useful AI projects happening here and there.

In March, DeepMind announced that it hadused a machine-learning technique called "free modelling" to detail the structures of six proteins associated with SARS-CoV-2, the coronavirus that causes the Covid-19 disease.Elsewhere, Israeli start-up Aidoc is using AI imaging to flag abnormalities in the lungs and a U.K. start-up founded by Viagra co-inventor David Brown is using AI to look for Covid-19 drug treatments.

Verena Rieser, a computer science professor at Heriot-Watt University, pointed out that autonomous robots can be used to help disinfect hospitals and AI tutors can support parents with the burden of home schooling. She also said "AI companions" can help with self isolation, especially for the elderly.

"At the periphery you can imagine it doing some stuff with CCTV," said Lawrence, adding that cameras could be used to collect data on what percentage of people are wearing masks.

Separately, a facial recognition system built by U.K. firm SCC has also been adapted to spot coronavirus sufferers instead of terrorists.In Oxford, England, Exscientia is screening more than 15,000 drugs to see how effective they are as coronavirus treatments. The work is being done in partnership withDiamond Light Source, the U.K.'s national "synchotron."

But AI's role in this pandemic is likely to be more nuanced than some may have anticipated. AI isn't about to get us out of the woods any time soon.

"It's kind of indicating how hyped AI was," said Lawrence, who is now a professor of machine learning at the University of Cambridge. "The maturity of techniques is equivalent to the noughties internet."

AI researchers rely on vast amounts of nicely labeled data to train their algorithms, but right now there isn't enough reliable coronavirus data to do that.

"AI learns from large amounts of data which has been manually labeled a time consuming and expensive task," said Catherine Breslin, a machine learning consultant who used to work on Amazon Alexa.

"It also takes a lot of time to build, test and deploy AI in the real world. When the world changes, as it has done, the challenges with AI are going to be collecting enough data to learn from, and being able to build and deploy the technology quickly enough to have an impact."

Breslin agrees that AI technologies have a role to play. "However, they won't be a silver bullet," she said, adding that while they might not directly bring an end to the virus, they can make people's lives easier and more fun while they're in lockdown.

The AI community is thinking long and hard about how it can make itself more useful.

Last week, Facebook AI announced a number of partnerships with academics across the U.S.

Meanwhile, DeepMind's polymath leader Demis Hassabis is helping the Royal Society, the world's oldest independent scientific academy, on a new multidisciplinary project called DELVE (Data Evaluation and Learning for Viral Epidemics). Lawrence is also contributing.

See the original post here:

A.I. can't solve this: The coronavirus could be highlighting just how overhyped the industry is - CNBC

Machine learning has quickly gone from niche field to crucial component of innumerable software stacks, but that doesnt mean its easy. The tools needed to create and manage it are enterprise-grade and often enterprise-only but Determined AI aims to make them more accessible than ever by open-sourcing its entire AI infrastructure product.

The company created its Determined Training Platform for developing AI in an organized, reliable way the kind of thing that large companies have created (and kept) for themselves, the team explained when they raised an $11 million Series A last year.

Machine learning is going to be a big part of how software is developed going forward. But in order for companies like Google and Amazon to be productive, they had to build all this software infrastructure, said CEO Evan Sparks. One company we worked for had 70 people building their internal tools for AI. There just arent that many companies on the planet that can withstand an effort like that.

At smaller companies, ML is being experimented with by small teams using tools intended for academic work and individual research. To scale that up to dozens of engineers developing a real product there arent a lot of options.

Theyre using things like TensorFlow and PyTorch, said Chief Scientist Ameet Talwalkar. A lot of the way that work is done is just conventions: How do the models get trained? Where do I write down the data on which is best? How do I transform data to a good format? All these are bread and butter tasks. Theres tech to do it, but its really the Wild West. And the amount of work you have to do to get it set up theres a reason big tech companies build out these internal infrastructures.

Determined AI, whose founders started out at UC Berkeleys AmpLab (home of Apache Spark), has been developing its platform for a few years, with feedback and validation from some paying customers. Now, they say, its ready for its open source debut with an Apache 2.0 license, of course.

We have confidence people can pick it up and use it on their own without a lot of hand-holding, said Sparks.

You can spin up your own self-hosted installation of the platform using local or cloud hardware, but the easiest way to go about it is probably the cloud-managed version that automatically provisions resources from AWS or wherever you prefer and tears them down when theyre no longer needed.

The hope is that the Determined AI platform becomes something of a base layer that lots of small companies can agree on, providing portability to results and standards so youre not starting from scratch at every company or project.

With machine learning development expected to expand by orders of magnitude in the coming years, even a small piece of the pie is worth claiming, but with luck, Determined AI may grow to be the new de facto standard for AI development in small and medium businesses.

You can check out the platform on GitHub or at Determined AIs developer site.

More here:

Determined AI makes its machine learning infrastructure free and open source - TechCrunch

German sportswear giant Adidas plans to bolster its lagging performance by shifting strategy to focus on its digital platform.

The fitness brand reported a 19% sales decline this week as more than 70% of its global stores remained closed during the coronavirus pandemic.

Over an earnings call with company executives, Adidas zeroed in on plans for "digital acceleration" and said it would pivot to a greater focus on its digital presence now that it has limited capacity to operate out of physical stores.

Business Insider got some insights from Adidas into what its revamped digital focus involves:

Adilette sales, a favorite among workers from home, went up by a 'triple-digit percentage' rate in April. Adidas

Adidas also says it is investing in its digital retail infrastructure, and is reaping the benefits during the coronavirus as its online stores, as well of those of partner retailers, are open all the time. Its physical stores and other sporting-goods retailers remain closed in most parts of the world.

The brand's ecommerce sales saw the highest growth rate on record last year, a jump of 34%. That spurt accelerated even further as sales for ecommerce grew by 35% in currency-neutral terms with 55% growth in March this year.

The firm aims to support this trend by connecting experts from its retail teams, mobile-fitness app Runtastic, and the IT team, all working towards building up ecommerce. Adidas is also extending partnership programs with digital pure players companies that only run virtually such as Zalando, Asos, Zappos, and Tmall.

Adidas' digital revamp is an extension of its 2015 'Creating the New' strategy. Adidas

As well as focusing more heavily on ecommerce, Adidas is also promoting heavily products that are selling well while people are stuck at home. The brand is increasingly promoting products like the Adilette Slides a favorite among consumers working from home as they have enjoyed high demand on the brand's channels. Adilette sales went up by a 'triple-digit percentage' rate in April, the company said.

For its sports apps "adidas Running" and "adidas Training", the brand has been offering free premium access since the start of the coronavirus pandemic. Since then, hundreds of thousands of athletes have used more than 250 training videos, workouts and training plans to continue keeping fit from home.

Adidas has also focused its marketing investments and efforts towards digital and social media channels and an increase in digital storytelling under the hashtag "#hometeam."

Over the past few weeks, the brand has shared inspiring home stories from the daily lives of numerous brand ambassadors around the world who demonstrate creative ways to "make the most of time at home and, of course, to continue to exercise."

Read more:

Adidas turns to AI and 3D printing to boost sales during coronavirus - Business Insider - Business Insider

The coronavirus outbreak is causing the world economy to go into a deep recession. Millions of people are expected to be laid off from their current jobs and re-enter the job market due to the economic downturn caused by the COVID-19. They need to know where their skills stand and whether they would have to transfer to new jobs. This is where Ashutosh Garg's and Varun Kochalia's Eightfold.ai steps in to make employees relevant in the post-COVID world.

Founded in 2016, Eightfold uses deep learning and artificial intelligence to break down peoples work into skills and other transferable experiences. It can match internal employees to new opportunities to put their skills to work. It can help recruiters find people with the potential to fill their openings, even if theyre not in a job with a matching job title right now. And it can provide job seekers much more visibility into how they can fit into a company with hundreds of job openings.

Ashutosh Garg, CEO of Eightfold

Ashutosh Garg graduated from IIT-Delhi in 1997 before he went to the US to finish his PhD in Computer Engineering at the University of Illinois, Urbana-Champaign. He went on to work with IBM Research. He has several thousand research citations and more than fifty patents.

Varun Kacholia graduated from IIT-Bombay in 2004 and finished his Master's from Berkeley. Proficient in search, ranking, and machine-learning, he led the News Feed team at Facebook, and the YouTube Search and Recommendations team at Google.

The founders met at Google and over time, realised that they could apply their AI experience to solving a new challenge where AI solutions could make a real difference to employment. The founders decided to go after the challenge of finding talent and providing the best career options.

Eightfold wants to help people achieve the best career they can, freed up from the outdated over-reliance on tools and processes like resumes, bad job interviews, and job descriptions.

Eightfold recognised the problems faced by employees:

1) People quit their jobs for new challenges and new roles when there are plenty of challenges in their own companies to apply their skills to new roles;

2) Recruiters get hundreds of resumes, but dont have a great way of seeing whos a goodmatch for a job.

3) Job seekers dont have a good way of searching for a job; they are stuck searching by keywords and titles instead of seeing how they have the potential to apply their skills to new challenges. There is very little hiring for potential, which hurts both companies and job seekers.

Eightfold.ai has partners, customers, and a full-time office in India supporting their B2B product. They see the market in India and Asia as huge for AI as companies need a better way of matching the right people to right jobs, whether new candidates or internal employees. There is also the need to improve retention rates, succession planning, and internal mobility.

"Eightfolds Talent Intelligence Platform matches individuals to opportunities at scale. The Talent Intelligence Platform has analysed more than one billion career paths, and calculated the individual factors such as experiences, education, and skills that predict movement through career paths. These calculations form a model that is applied to the career path of any individual and any position. For an individual, the model will rank available positions and show the individual why a position is a match for their qualifications. For a position, the model will rank available candidates and show a recruiter or hiring manager why a person is a match for the position requirements," says Ashutosh.

Eightfold says the Talent Intelligence Platform is able to match individuals and opportunities based on their potential to succeed in a role, rather than on past success in an equivalent role. The AI helps recogniseemployees who might feel stale in their current positions, provide themwith job openings that match their skills and ambitions.

Eightfold is saving companies time and money. The platform reduces the time it takes to hire someone, honing in on the most qualified candidates, and cutting down on the need for a lot of recruitment agencies and other spendings. With this platform, AI can surface many candidates who are already in a companys network: past applicants, current employees, alumni, and employee referrals, for example.

He adds that for Eightfold, diversity and inclusion are a major part of their mission, rather than an afterthought. The platform allows the individual to mask a persons ethnicity or gender, thus stripping away any unconscious bias when a slate of candidates is delivered to the manager.

Since they are a software-as-a-service (SaaS) company, they sell the platform as asolution to businesses on an annual subscription basis.

Their first customer was Tata Communications, with more than 10,000 global employees. It has been using Eightfold to free itself up from the time-intensive and redundant parts of building a talent pipeline, and concentrate instead on creating the best experience possible for both candidates and hiring managers.

"We signed up Tata Communications as our first customer more than three years ago because of the potential of our Talent Intelligence Platform to address several of its talent needs with matching technology," he explains.

They have 100-plus customers all over the world.

Changing the humanresources field, which has a lot of outdated processes and technologies, is a considerable task. One of the challenges they face is to show customers how their approach with AI can have a positive impact on many aspects of talent relationships.

"Everyones challenges have a lot in common. Theyre not that different. Large organisationshave too many of the wrong applicants. They dont have good ways of screening. Theyre losing good people who want new challenges, and could have metthose challenges internally. These concerns may vary somewhat by country, industry, and by the specific company, but overall they are not very different everywhere we go," says Ashutosh.

The company has raised $55 million in Series-C. They compete with Humantelligence, Crosschq, Searchlight, and Fintros.

According to Gartner, SaaS is a $150 billion opportunity. According to SaaS Boomi, India has more than 250 companies with more than $1 million in revenues.

How has the coronavirus outbreak disrupted your life? And how are you dealing with it? Write to us or send us a video with subject line 'Coronavirus Disruption' to editorial@yourstory.com

Read the original:

This startup by San Francisco-based founders uses AI to find the right hires for corporates - YourStory

Ive spent a good chunk of my workday so far creating memes thanks to this amazing website from Imgflip that automatically generates captions for memes using a neural network. Im addicted because the site A) takes the pressure off trying to be clever by auto-filling the captions; B) actually, somehow, regularly generates clever captions; and C) sometimes creates captions that make no sense, which are hilarious anyway.

You can pick from 48 classic meme templates, including distracted boyfriend, Drake in Hotline Bling, mocking Spongebob, surprised Pikachu, and Oprah giving things away. To generate meme captions, you just have to click on the meme template on the top of the page. If you dont like what the site serves up, or you just want to see what other ridiculous caption you might get, you can click a refresh button to get a new caption for the same meme. If you want to use an entirely different meme image, you just have to click over to another one.

The neural network was trained using public images uploaded to the sites meme maker, according to Imgflip. The company warns that no profanity filtering was done on the training data, so theres a chance you might see inappropriate language in the auto-generated captions.

I mobilized some colleagues to make some memes, and Ive collected a few of my favorites below. You can also see a list of popular memes made by the generator right here.

Happy memeing!

Read the original:

Neural net-generated memes are one of the best uses of AI on the internet - The Verge

Virtual training is more important than ever. This release gives Rehearsal users a platform that is scalable beyond anything available today. New functionality puts learners in the drivers seat while ensuring their efforts are targeted and effective, said Darik Volpa, founder and CEO of Rehearsal

RENO, Nev. (PRWEB) May 04, 2020

Today Rehearsal, the provider of best-in-class video-based practice software and services, announced a major release that extensively increases the scalability of the platform. This is achieved through a learner-driven approach that utilizes self-guided practice and artificial intelligence (AI). This release significantly reduces mentor workload, which is a frequent bottleneck in skills development. Learners using the platform receive AI-driven keyword analysis, are prompted to evaluate their own submissions using a customizable rubric, and experience accelerated skills development by automatically completing selected assignment activities.

Virtual training is more important than ever. This release gives Rehearsal users a platform that is scalable beyond anything available today. New functionality puts learners in the drivers seat while ensuring their efforts are targeted and effective, said Darik Volpa, founder and CEO of Rehearsal. By removing the mentor bottleneck, Rehearsal can be implemented at a scale previously unimaginable. This release further strengthens our position as we deliver the most scalable video-based practice and coaching platform on the market.

Prior to this release, Rehearsal assignment activities required mentor involvement in each learner submission, resulting in programs that were limited by mentor bandwidth. With the addition of AI to the Rehearsal platform, mentor bandwidth no longer limits the size and scope of programs. Learners now benefit from clear and highly objective feedback, making their video-based practice and coaching experience more empowering and effective.

Leah Barrett, Product Manager at Rehearsal mentioned, The shift towards a more learner-centric experience using AI is just the beginning of what we are calling an Intelligent Journey, where learning is self-guided, targeted, and scalable. Future releases will reinforce this as we maintain an intense focus on further advancing the industrys best video-based practice platform.

This release is available for both web and mobile applications with full functionality. This ensures that learners, mentors, and program administrators have convenient access and capabilities anytime, anywhere, and on any device. If needed, traditional Rehearsal functionality through which mentors provide tailored one-on-one feedback and guidance is still conveniently accessible. This versatility makes the platform a valuable solution for a wide number of use-cases across many industry verticals.

For more information, please visit: https://www.rehearsal.com/campaigns/scale-training-with-self-guided-skill-development/?r=IJPR20.

###

About RehearsalRehearsal is a collaborative video-based practice platform. It gives your team a safe place to develop sales and communication skills so they can perform when it matters. Conveniently practice, coach, and collaborate to inspire the whole team. Rehearsal was named a leader in the Sales Coaching and Learning market by Aragon Research, featured on the Selling Power 2018 List of Recommended Sales Enablement Partners and awarded a 2017 Brandon Hall Group Excellence in Learning Awards.

Share article on social media or email:

Go here to see the original:

Rehearsal Announces Major Release Incorporating Self-Guided Practice and Artificial Intelligence - PR Web

The search for a COVID-19 vaccine is on. Hopefully, advancements in artificial intelligence will make the time required to produce a vaccine shorter.

Artificial intelligence is revolutionizing the pharmaceutical industry by empowering companies to dramatically accelerate the development of new drugs while cutting costs by hundreds of millions of dollars.

But the growing use of the technology in the industry is also fast creating new risks and updating old ones.

By enabling scientists to analyze millions of data points with a few clicks of a mouse, AI technologies such as natural language processing and machine learning have the potential to slash the time it takes for companies to comply with the scientific and regulatory requirements necessary to research, test and market new drugs by several years.

This kind of efficiency-boosting potential comes in handy as the industry is pushed by governments and by public opinion to devise quick solutions to the devastating COVID-19 crisis.

The widespread use of highthroughput screening in the lab, namely, the technique of testing large collections of compounds using robotics, coupled with AI, will speed up the testing of vaccine candidates to detect an immune response, said Matthew Clark, the director of Global Markets at brokers La Playa.

The use of computational methods to track and record clinical trials data, and seek out trends in it, will surely speed the process still further.

While it can take up to eight or 10 years and hundreds of millions and even a few billion dollars to research, test and start manufacturing a new drug with traditional methods, AI can significantly streamline the process and reduce costs.

The importance of a fast-tracked process to the bottom line of pharma companies is punctuated by estimates that around 90% of the drugs that reach initial trial phases end up failing.

As many companies are focusing on COVID-19 right now, I believe that it will take one to two years to find a treatment for it, said Nagesh Jadhav, a digital transformation and innovation leader at consultants Stefanini.

The pharmaceutical industry is accumulating a huge amount of data about microorganisms and genome sequencing of drug candidates. There are thousands and thousands of drug candidates and it becomes very hard to manually screen them.

Large groups, such as GlaxoSmithKline, have announced partnerships with AI tech firms to fast-track COVID-19 treatments.

In April, UK-based biotech firm Healx announced it was employing AI to investigating treatments for COVID-19 by looking at combinations of 4,000 different medications available in the market. That implies analyzing 8 million pairs and 10.5 billion drug triples, according to the company.

The sheer amount of work to assess the combinations would be enough to keep a team of scientists busy for a long time. With AI, however, the expectation is that valid conclusions will be reached in a much shorter time.

Successful examples of the application of the technology have already popped up.

In January, Sumitomo Dainippon Pharma, a Japanese drug maker, and Exscientia, a British AI developer, announced they had created a new treatment for obsessive compulsive disorder, making the drug available for phase 1 clinical trials after less than 12 months of AI-based research. Usually, the same work would have taken around four to five years, the companies estimate.

Tom Daniels, life science practice leader, HDI Global SE

Clinical trials can also benefit from the technology, noted Tom Daniels, the life science practice leader at HDI Global SE.

AI has the potential to match suitable patients, based on factors such as age, medical history, location or symptoms, to a clinical trial they may be eligible for, thereby streamlining the recruitment and screening phases of a trial, he said.

There is also the potential for AI-enabled sensors to be used during a trial to ensure adherence to the study protocol by sensing whether a dose has been missed and issuing an alert. This can ensure trial integrity, improved accuracy and help to explain anomalies in the trial data, which will be presented to the regulator if applying for approval.

Machine learning has also enabled the industry to collate medical information sourced from scientific literature, as scientists rush to find a cure for COVID-19.

One company in particular used its system to search for approved drugs that had the potential to reduce the ability of COVID-19 to infect lung cells, noted Tanya Patel, an underwriter with HDI Global SE.

This list was whittled down, based upon criteria such as side effects and required dosage, to identify the most promising candidate, an approved rheumatoid arthritis drug, which can now be trialed, she said.

At least 11 other molecules have been identified through AI technology as potential candidates by various firms globally.

But deploying powerful new technologies often implies new risks for companies, and the marriage between pharma and AI is no exception to the rule.

Not the least because, to a certain extent, drug makers are treading into a realm that is not really their strongest suit.

The core expertise of a biopharma company is not AI on itself. Many times they have to contract out a third party product or service as the starting point for how they are going to apply AI to their product development, said Brad Johns, the head of life sciences at The Hartford.

An algorithm can come with bugs, issues or biases that pharma companies are not aware of.

As other information-based technologies, the quality of the work performed by AI-based systems is only as good as the data fed into them. In an accelerated R&D and trial process, the usual controls that pharma companies employ to find problems in the development of drugs may not be sufficient to spot them on time.

Narvin Powar, director, PharmaInsure

With AI, companies can potentially develop drugs at a much faster pace. As a result, errors and omissions may not be spotted until much later at the end of the process, said Narvin Powar, a director at broker PharmaInsure.

Privacy concerns related to patient data are another worry, as the handling of enormous amounts of information from public health systems, hospital databases and clinical trials raise ethical and regulatory issues.

Pharma firms are a trove of riches for hackers, and if these firms are hacked, they can find themselves in violation of regulations such as the U.S. Health Insurance Portability and Accountability Act or Europes General Data Protection Regulations, which can result in large fines and reputational damage.

There is no consistent regulatory framework on how companies must use peoples information, Johns pointed out. What a company does may be acceptable and completely compliant in one jurisdiction, but not in other.

And then there is the matter of liability, a big concern for pharma companies in any situation, but particularly when factoring in the risk added by partners that work in areas out of the industrys realm of expertise.

There may be little clarity about where the liability of the pharmaceutical company stops and where the liability of the AI technology partner starts, Powar said.

Although it is possible to deal with it in contracts, there will always be gray areas.

This is exactly the kind of risk that gets exacerbated during a time of emergency.

In recent months, companies have been pushed to come up with solutions for COVID-19, as media personalities and even presidents have touted the potential of unproven medicines that, if widely used without proper due diligence, could result in lawsuits for their producers.

Despite the obvious pressure to get treatments to patients as quickly as possible, it is still necessary to properly test or assess any vaccine to ensure its safe and effective, Clark said.

As ever, regulators have a key role to play to ensure corners can be cut without compromising patient safety.&

Read more:

How AI Can Help Pharma Risk Managers Beat Their Regulatory Nightmares and Get the Rest They Need - Workers Comp Forum

The pandemic, in other words, has turned into a gateway for AI adoption in health carebringing both opportunity and risk. On the one hand, it is pushing doctors and hospitals to fast-track promising new technologies. On the other, this accelerated process could allow unvetted tools to bypass regulatory processes, putting patients in harms way.

At a high level, artificial intelligence in health care is very exciting, says Chris Longhurst, the chief information officer at UC San Diego Health. But health care is one of those industries where there are a lot of factors that come into play. A change in the system can have potentially fatal unintended consequences.

Before the pandemic, health-care AI was already a booming area of research. Deep learning, in particular, has demonstrated impressive results for analyzing medical images to identify diseases like breast and lung cancer or glaucoma at least as accurately as human specialists. Studies have also shown the potential of using computer vision to monitor elderly people in their homes and patients in intensive care units.

But there have been significant obstacles to translating that research into real-world applications. Privacy concerns make it challenging to collect enough data for training algorithms; issues related to bias and generalizability make regulators cautious to grant approvals. Even for applications that do get certified, hospitals rightly have their own intensive vetting procedures and established protocols. Physicians, like everybody elsewere all creatures of habit, says Albert Hsiao, a radiologist at UCSD Health who is now trialing his own covid detection algorithm based on chest x-rays. We dont change unless were forced to change.

As a result, AI has been slow to gain a foothold. It feels like theres something there; there are a lot of papers that show a lot of promise, said Andrew Ng, a leading AI practitioner, in a recent webinar on its applications in medicine. But its not yet as widely deployed as we wish.

QURE.AI

Pierre Durand, a physician and radiologist based in France, experienced the same difficulty when he cofounded the teleradiology firm Vizyon in 2018. The company operates as a middleman: it licenses software from firms like Qure.ai and a Seoul-based startup called Lunit and offers the package of options to hospitals. Before the pandemic, however, it struggled to gain traction. Customers were interested in the artificial-intelligence application for imaging, Durand says, but they could not find the right place for it in their clinical setup.

The onset of covid-19 changed that. In France, as caseloads began to overwhelm the health-care system and the government failed to ramp up testing capacity, triaging patients via chest x-raythough less accurate than a PCR diagnosticbecame a fallback solution. Even for patients who could get genetic tests, results could take at least 12 hours and sometimes days to returntoo long for a doctor to wait before deciding whether to isolate someone. By comparison, Vizyons system using Lunits software, for example, takes only 10 minutes to scan a patient and calculate a probability of infection. (Lunit says its own preliminary study found that the tool was comparable to a human radiologist in its risk analysis, but this research has not been published.) When there are a lot of patients coming, Durand says, its really an attractive solution.

Vizyon has since signed partnerships with two of the largest hospitals in the country and says it is in talks with hospitals in the Middle East and Africa. Qure.ai, meanwhile, has now expanded to Italy, the US, and Mexico on top of existing clients. Lunit is also now working with four new hospitals each in France, Italy, Mexico, and Portugal.

In addition to the speed of evaluation, Durand identifies something else that may have encouraged hospitals to adopt AI during the pandemic: they are thinking about how to prepare for the inevitable staff shortages that will arise after the crisis. Traumatic events like a pandemic are often followed by an exodus of doctors and nurses. Some doctors may want to change their way of life, he says. Whats coming, we dont know.

Hospitals new openness to AI tools hasnt gone unnoticed. Many companies have begun offering their products for a free trial period, hoping it will lead to a longer contract.

It's a good way for us to demonstrate the utility of AI, says Brandon Suh, the CEO of Lunit. Prashant Warier, the CEO and cofounder of Qure.ai, echoes that sentiment. In my experience outside of covid, once people start using our algorithms, they never stop, he says.

Both Qure.ais and Lunits lung screening products were certified by the European Unions health and safety agency before the crisis. In adapting the tools to covid, the companies repurposed the same functionalities that had already been approved.

QURE.AI

Qure.ais qXR, for example, uses a combination of deep-learning models to detect common types of lung abnormalities. To retool it, the firm worked with a panel of experts to review the latest medical literature and determine the typical features of covid-induced pneumonia, such as opaque patches in the image that have a ground glass pattern and dense regions on the sides of the lungs. It then encoded that knowledge into qXR, allowing the tool to calculate the risk of infection from the number of telltale characteristics present in a scan. A preliminary validation study the firm ran on over 11,000 patient images found that the tool was able to distinguish between covid and non-covid patients with 95% accuracy.

But not all firms have been as rigorous. In the early days of the crisis, Malik exchanged emails with 36 companies and spoke with 24, all pitching him AI-based covid screening tools. Most of them were utter junk, he says. They were trying to capitalize on the panic and anxiety. The trend makes him worry: hospitals in the thick of the crisis may not have time to perform due diligence. When youre drowning so much, he says, a thirsty man will reach out for any source of water.

Kay Firth-Butterfield, the head of AI and machine learning at the World Economic Forum, urges hospitals not to weaken their regulatory protocols or formalize long-term contracts without proper validation. Using AI to help with this pandemic is obviously a great thing to be doing, she says. But the problems that come with AI dont go away just because there is a pandemic.

UCSDs Longhurst also encourages hospitals to use this opportunity to partner with firms on clinical trials. We need to have clear, hard evidence before we declare this as the standard of care, he says. Anything less would be a disservice to patients.

To have more stories like this delivered directly to your inbox,sign upfor our Webby-nominated AI newsletter The Algorithm. It's free.

Read more:

Doctors are using AI to triage covid-19 patients. The tools may be here to stay - MIT Technology Review

Artificial intelligence has become a focus of certain ethical concerns, but it also has some major sustainability issues.

Last June, researchers at the University of Massachusetts at Amherst released a startling report estimating that the amount of power required for training and searching a certain neural network architecture involves the emissions of roughly 626,000 pounds of carbon dioxide. Thats equivalent to nearly five times the lifetime emissions of the average U.S. car, including its manufacturing.

This issue gets even more severe in the model deployment phase, where deep neural networks need to be deployed on diverse hardware platforms, each with different properties and computational resources.

MIT researchers have developed a new automated AI system for training and running certain neural networks. Results indicate that, by improving the computational efficiency of the system in some key ways, the system can cut down the pounds of carbon emissions involved in some cases, down to low triple digits.

The researchers system, which they call a once-for-all network, trains one large neural network comprising many pretrained subnetworks of different sizes that can be tailored to diverse hardware platforms without retraining. This dramatically reduces the energy usually required to train each specialized neural network for new platforms which can include billions of internet of things (IoT) devices. Using the system to train a computer-vision model, they estimated that the process required roughly 1/1,300 the carbon emissions compared to todays state-of-the-art neural architecture search approaches, while reducing the inference time by 1.5-2.6 times.

The aim is smaller, greener neural networks, says Song Han, an assistant professor in the Department of Electrical Engineering and Computer Science. Searching efficient neural network architectures has until now had a huge carbon footprint. But we reduced that footprint by orders of magnitude with these new methods.

The work was carried out on Satori, an efficient computing cluster donated to MIT by IBM that is capable of performing 2 quadrillion calculations per second. The paper is being presented next week at the International Conference on Learning Representations. Joining Han on the paper are four undergraduate and graduate students from EECS, MIT-IBM Watson AI Lab, and Shanghai Jiao Tong University.

Creating a once-for-all network

The researchers built the system on a recent AI advance called AutoML (for automatic machine learning), which eliminates manual network design. Neural networks automatically search massive design spaces for network architectures tailored, for instance, to specific hardware platforms. But theres still a training efficiency issue: Each model has to be selected then trained from scratch for its platform architecture.

How do we train all those networks efficiently for such a broad spectrum of devices from a $10 IoT device to a $600 smartphone? Given the diversity of IoT devices, the computation cost of neural architecture search will explode, Han says.

The researchers invented an AutoML system that trains only a single, large once-for-all (OFA) network that serves as a mother network, nesting an extremely high number of subnetworks that are sparsely activated from the mother network. OFA shares all its learned weights with all subnetworks meaning they come essentially pretrained. Thus, each subnetwork can operate independently at inference time without retraining.

The team trained an OFA convolutional neural network (CNN) commonly used for image-processing tasks with versatile architectural configurations, including different numbers of layers and neurons, diverse filter sizes, and diverse input image resolutions. Given a specific platform, the system uses the OFA as the search space to find the best subnetwork based on the accuracy and latency tradeoffs that correlate to the platforms power and speed limits. For an IoT device, for instance, the system will find a smaller subnetwork. For smartphones, it will select larger subnetworks, but with different structures depending on individual battery lifetimes and computation resources. OFA decouples model training and architecture search, and spreads the one-time training cost across many inference hardware platforms and resource constraints.

This relies on a progressive shrinking algorithm that efficiently trains the OFA network to support all of the subnetworks simultaneously. It starts with training the full network with the maximum size, then progressively shrinks the sizes of the network to include smaller subnetworks. Smaller subnetworks are trained with the help of large subnetworks to grow together. In the end, all of the subnetworks with different sizes are supported, allowing fast specialization based on the platforms power and speed limits. It supports many hardware devices with zero training cost when adding a new device.In total, one OFA, the researchers found, can comprise more than 10 quintillion thats a 1 followed by 19 zeroes architectural settings, covering probably all platforms ever needed. But training the OFA and searching it ends up being far more efficient than spending hours training each neural network per platform. Moreover, OFA does not compromise accuracy or inference efficiency. Instead, it provides state-of-the-art ImageNet accuracy on mobile devices. And, compared with state-of-the-art industry-leading CNN models , the researchers say OFA provides 1.5-2.6 times speedup, with superior accuracy. Thats a breakthrough technology, Han says. If we want to run powerful AI on consumer devices, we have to figure out how to shrink AI down to size.

The model is really compact. I am very excited to see OFA can keep pushing the boundary of efficient deep learning on edge devices, says Chuang Gan, a researcher at the MIT-IBM Watson AI Lab and co-author of the paper.

If rapid progress in AI is to continue, we need to reduce its environmental impact, says John Cohn, an IBM fellow and member of the MIT-IBM Watson AI Lab. The upside of developing methods to make AI models smaller and more efficient is that the models may also perform better.

Read the original post:

Reducing the carbon footprint of artificial intelligence - MIT News