Cloud Hosting

Companies come in promising the world and often dont deliver, said Bob Wachter, head of the department of medicine at the University of California, San Francisco. When I look for examples of true AI and machine learning thats really making a difference, theyre pretty few and far between. Its pretty underwhelming.

Administrators say algorithms the software that processes data from outside companies dont always work as advertised because each health system has its own technological framework. So hospitals are building out engineering teams and developing artificial intelligence and other technology tailored to their own needs.

But its slow going. Research based on job postings shows health care behind every industry except construction in adopting AI.

The Food and Drug Administration has taken steps to develop a model for evaluating AI, but it is still in its early days. There are questions about how regulators can monitor algorithms as they evolve and rein in the technologys detrimental aspects, such as bias that threaten to exacerbate health care inequities.

Sometimes theres an assumption that AI is working, and its just a matter of adopting it, which is not necessarily true, said Florenta Teodoridis, a professor at the University of Southern Californias business school whose research focuses on AI. She added that being unable to understand why an algorithm came to a certain result is fine for things like predicting the weather. But in health care, its impact is potentially life-changing.

Despite the obstacles, the tech industry is still enthusiastic about AIs potential to transform health care.

The transition is slightly slower than I hoped but well on track for AI to be better than most radiologists at interpreting many different types of medical images by 2026, Hinton told POLITICO via email. He said he never suggested that we should get rid of radiologists, but that we should let AI read scans for them.

If hes right, artificial intelligence will start taking on more of the rote tasks in medicine, giving doctors more time to spend with patients to reach the right diagnosis or develop a comprehensive treatment plan.

I see us moving as a medical community to a better understanding of what it can and cannot do, said Lara Jehi, chief research information officer for the Cleveland Clinic. It is not going to replace radiologists, and it shouldnt replace radiologists.

Radiology is one of the most promising use cases for AI. The Mayo Clinic has a clinical trial evaluating an algorithm that aims to reduce the hours-long process oncologists and physicists undertake to map out a surgical plan for removing complicated head and neck tumors.

An algorithm can do the job in an hour, said John D. Halamka, president of Mayo Clinic Platform: Weve taken 80 percent of the human effort out of it. The technology gives doctors a blueprint they can review and tweak without having to do the basic physics themselves, he said.

NYU Langone Health has also experimented with using AI in radiology. The health system has collaborated with Facebooks Artificial Intelligence Research group to reduce the time it takes to get an MRI from one hour to 15 minutes. Daniel Sodickson, a radiological imaging expert at NYU Langone who worked on the research, sees opportunity in AIs ability to downsize the amount of data doctors need to review.

When I look for examples of true AI and machine learning thats really making a difference, theyre pretty few and far between. Its pretty underwhelming.

Bob Wachter, head of the department of medicine at the University of California, San Francisco

Covid has accelerated AIs development. Throughout the pandemic, health providers and researchers shared data on the disease and anonymized patient data to crowdsource treatments.

Microsoft and Adaptive Biotechnologies, which partner on machine learning to better understand the immune system, put their technology to work on patient data to see how the virus affected the immune system.

The amount of knowledge thats been obtained and the amount of progress has just been really exciting, said Peter Lee, corporate vice president of research and incubations at Microsoft.

There are other success stories. For example, Ochsner Health in Louisiana built an AI model for detecting early signs of sepsis, a life-threatening response to infection. To convince nurses to adopt it, the health system created a response team to monitor the technology for alerts and take action when needed.

Im calling it our care traffic control, said Denise Basow, chief digital officer at Ochsner Health. Since implementation, she said, death from sepsis is declining.

The biggest barrier to the use of artificial intelligence in health care has to do with infrastructure.

Health systems need to enable algorithms to access patient data. Over the last several years, large, well-funded systems have invested in moving their data into the cloud, creating vast data lakes ready to be consumed by artificial intelligence. But thats not as easy for smaller players.

Another problem is that every health system is unique in its technology and the way it treats patients. That means an algorithm may not work as well everywhere.

Over the last year, an independent study on a widely used sepsis detection algorithm from EHR giant Epic showed poor results in real-world settings, suggesting where and how hospitals used the AI mattered.

This quandary has led top health systems to build out their own engineering teams and develop AI in-house.

That could create complications down the road. Unless health systems sell their technology, its unlikely to undergo the type of vetting that commercial software would. That could allow flaws to go unfixed for longer than they might otherwise. Its not just that the health systems are implementing AI while no ones looking. Its also that the stakeholders in artificial intelligence, in health care, technology and government, havent agreed upon standards.

A lack of quality data which gives algorithms material to work with is another significant barrier in rolling out the technology in health care settings.

Over the last several years, large, well-funded systems have invested in moving their data into the cloud, creating vast data lakes ready to be consumed by artificial intelligence.|Elaine Thompson/AP Photo

Much data comes from electronic health records but is often siloed among health care systems, making it more difficult to gather sizable data sets. For example, a hospital may have complete data on one visit, but the rest of a patients medical history is kept elsewhere, making it harder to draw inferences about how to proceed in caring for the patient.

We have pieces and parts, but not the whole, said Aneesh Chopra, who served as the governments chief technology officer under former President Barack Obama and is now president of data company CareJourney.

While some health systems have invested in pulling data from a variety of sources into a single repository, not all hospitals have the resources to do that.

Health care also has strong privacy protections that limit the amount and type of data tech companies can collect, leaving the sector behind others in terms of algorithmic horsepower.

Importantly, not enough strong data on health outcomes is available, making it more difficult for providers to use AI to improve how they treat patients.

That may be changing. A recent series of studies on a sepsis algorithm included copious details on how to use the technology in practice and documented physician adoption rates. Experts have hailed the studies as a good template for how future AI studies should be conducted.

But working with health care data is also more difficult than in other sectors because it is highly individualized.

We found that even internally across our different locations and sites, these models dont have a uniform performance, said Jehi of the Cleveland Clinic.

And the stakes are high if things go wrong. The number of paths that patients can take are very different than the number of paths that I can take when Im on Amazon trying to order a product, Wachter said.

Health experts also worry that algorithms could amplify bias and health care disparities.

For example, a 2019 study found that a hospital algorithm more often pushed white patients toward programs aiming to provide better care than Black patients, even while controlling for the level of sickness.

Last year, the FDA published a set of guidelines for using AI as a medical device, calling for the establishment of good machine learning practices, oversight of how algorithms behave in real-world scenarios and development of research methods for rooting out bias.

The agency subsequently published more specific guidelines on machine learning in radiological devices, requiring companies to outline how the technology is supposed to perform and provide evidence that it works as intended. The FDA has cleared more than 300 AI-enabled devices, largely in radiology, since 1997.

Regulating algorithms is a challenge, particularly given how quickly the technology advances. The FDA is attempting to head that off by requiring companies to institute real-time monitoring and submit plans on future changes.

But in-house AI isnt subject to FDA oversight. Bakul Patel, former head of the FDAs Center for Devices and Radiological Health and now Googles senior director for global digital health strategy and regulatory affairs, said that the FDA is thinking about how it might regulate noncommercial artificial intelligence inside of health systems, but he adds, theres no easy answer.

FDA has to thread the needle between taking enough action to mitigate flaws in algorithms while also not stifling AIs potential, he said.

Some argue that public-private standards for AI would help advance the technology. Groups, including the Coalition for Health AI, whose members include major health systems and universities as well as Google and Microsoft, are working on this approach.

But the standards they envision would be voluntary, which could blunt their impact if not widely adopted.

See original here:
Artificial intelligence was supposed to transform health care. It hasn't. - POLITICO

QuantumBlack, AI by McKinsey recently sat down with Selim Turki, head of data and AI at Uber-owned mobility company Careem, to discuss the latest trends in advanced analytics and artificial intelligence. Far from a dry discussion of theory, the conversation coalesced around several fascinating use cases in which Careem is using AI to make a difference in peoples lives. We discussed how AI is being leveraged to improve customer and driver security through targeted facial-recognition checks to ensure drivers (captains) are who they say they are. We also discussed how AI is being used to provide customers with the most accurate and up-to-date estimated times of arrival (ETAs) by factoring in a host of conditions, including local weather conditions, prayer times, and even iftar times during Ramadan. Along the way, we discussed what it means to be an AI first company and the outlook for AI techand talentin the region.

QuantumBlack: Was AI always an important part of Careems growth journey? How has AIs role evolved since Careems inception?

Selim Turki: We started our journey as a ride-hailing company booking journeys for corporate clients. We were initially booking cars manually, without a data server, before introducing more advanced systems to deliver more efficient, personalized experiences. Since day one, our mission has been to simplify and improve the lives of peopleparticularly our customers and captains. We quickly understood that maintaining high reliability for our dynamic marketplace 24/7 was a complex process that needed to be driven by instant decision making through continuous automation at scale.

We began processing real-time data, using algorithms and machine learning [ML] models to solve some of the core issues for our ride-hailing marketplace, including matching customers and captains efficiently, shaping our demand and supply via surge pricing, calculating accurate ETAs for our captains, and improving our maps and location search functionality.

Today, we are scaling the Middle East, North Africa, and Pakistan [MENAP] regions first super app. AI is in our DNA as we invest more in platform capabilities and team skill sets. Our hiring strategy is focused on growing a diverse team of data and machine learning scientists to build out our in-house experimentation and machine learning platforms.

QuantumBlack: Has the adoption of these new AI techniques changed the way Careem works to serve its customers? How has this affected business teams within the organization?

Selim Turki: We use several AI techniques depending on the type of service we offer in our super app. All of these techniques are directed at three particular needs:

We use AI to factor in prayer times, iftar time during Ramadan, and weather conditions to better predict the ETA accuracy of when the food will be delivered to our customers.

QuantumBlack: How many AI practitioners work with Careem today?

Selim Turki: We have dozens of AI and machine learning experts who are driving forward our strategy of being an AI-first company. Part of our plan is to educate the entire organization on the topic, inviting our engineers and business counterparts to use AI to solve some of their challenges. We have also designed a program dedicated to new college graduates to ensure future talent is up to date with the latest AI techniques and to encourage them to further develop their skills.

QuantumBlack: How do you integrate AI into your decision making now? How do you stay ahead of competition in the market?

Selim Turki: AI is part of Careems decision-making framework. We set quarterly goals to measure and assess the usage and impact of our ML models on the different business streams.

We use rigorous statistical methodologies, taking confounding effects into account, to accurately estimate the models impact on different areas of the business.

To help our data and AI teams stay on top of the changes happening in the industry, we have started collaborating with regional academic institutions to solve some of the most significant super-app challenges and to identify exciting new opportunities for AI innovation.

We publish our progress on the Careem engineering blog and invite third parties to collaborate with us on specific areas related to AI.

We also contribute to open-source data communities and offer our work to other AI and ML professionals.

QuantumBlack: Can you share a recent instance of how AI fundamentally changed the way Careem does business with its customers or captains?

Selim Turki: With any digital platform, fraudsters will look for loopholes to exploit, whether through creating fake-identity accounts or exploring ways to hijack open accounts. Our team uses advanced AI techniques focusing on the identity of users to detect and prevent losses stemming from fraud. One system we use, called Crazy Wall, uses a relational graph convolutional network to map different data points of a customers identity. It also identifies characteristics shared across different identities to detect and mass-block fraudulent patterns across customer or captain activities.

QuantumBlack: AI talent has been a key challenge for companies in the region. How have you dealt with the regions structural talent issues?

Selim Turki: The regions tech talent is growing rapidly, and its exciting to see more specialists choosing to come to the region to make an impact in some of the fastest-growing countries in the world. Its also exciting to see a growing number of local university graduates specializing in AI. Were fortunate to have attracted a strong community of AI talent both locally and from surrounding markets to Careem. Our teams are building tech across various areas, including e-commerce, technology-enabled logistics, maps, identity, and fintech. They can solve complex and meaningful challenges at scale thanks to Careems deep tech expertise, strong regulatory relationships, local presence, and increasingly specialized global teams that are structured to operate as autonomous start-ups. Our team of more than 400 engineers and developers are empowered to develop cutting-edge technology every day. Being a remote-first company allows us to attract talent from across the world who want to have an impact on the MENAP region. This means that the opportunities to gain new perspectives and solve complex, real-world challenges alongside talented peers are endless.

QuantumBlack: Do you think the talent-supply challenges are here to stay? What is your ambition for attracting cutting-edge AI practitioners to Careem in the next three to five years?

Selim Turki: As AI becomes more widely used across industries, the demand for specialists will continue to rise. We need to inspire the next generation of data and AI specialists to be curious and gain exposure to the workplace at an earlier age.

At Careem, we are focused on building an AI culture where opportunities to learn and thrive are fostered by adapting, mentoring, and sharing within our AI communities and beyond. We are also hoping to make AI more accessible to stakeholders across Careem with initiatives like no-code AI, where AI is accessible without existing coding skill sets, as well as partnerships with AI labs to democratize AI usage across the company.

QuantumBlack: How will AI specifically change the mobility space in MENAP? Are there any white spaces where MENAP companies could be global first movers?

Selim Turki: The global mobility space is at a very nascent phase, with considerable opportunities to solve using AI techniques. At Careem, we have the vision of creating an internet-like network to transport packages of atoms, like how the internet transports packets of bits, called the AtomNet.

The AtomNet provides an open-network platform that connects, manages, and routes multimode autonomous vehicles [AVs] to make transport ubiquitous. Similar to how packets can travel across multiple modalities of transport (Wi-Fi, DSL, cable, and fiber), packages on the AtomNet can travel in autonomous motorcycles, cars, vans, trucks, ships, drones, and airplanes. We foresee an AtomNet industry ecosystem with open package headers and protocols to allow package switching and efficient package mobility. With open protocols, coordination costs will drop significantly, and local, national, and international transport gaps will narrow over the years.

AtomNet will support Careems quick commerce, fulfillment centers, restaurants, groceries, dark stores, transportation, and cross-border commerce. We see the epicenter of AtomNet starting in the UAE due to its progressive regulation and culture of innovation.

QuantumBlack: AI is still in its nascency in the broader context of this region. How do you think this will change in the next five to ten years?

Selim Turki: A long and exciting journey is ahead of us in the wider Middle East. With the growing pace of technology, more and more regional corporations will use AI to enhance their products and offer a better experience to customers.

At Careem, our primary focus will continue to be building the internet platform of the Middle East to provide access to our servicesusing data and AI as a core to simplify and improve customers lives. The meta goal is to delight all our users and personalize their experience through data and AI in every service offered through our super app.

The current trend of making trade-offs by improving AI prediction will be strengthened at the cost of short-term factors such as ingestion costs, customer experience, and operational excellence. We will continue investing in our data streams to help our models learn, build, and manage algorithms at scale. Moreover, real-time feedback loops will continue to decipher customer behavior and how it evolves by using our services through leveraging more intelligent software and hardware. Some of the emerging machine learning models will be tailored more to our region, considering language, customer behavior, and product relevance.

Our goal is to provide the simplest and best possible customer experience. To make things simple, you have to make them intuitive. To make things intuitively simple, we need to:

View original post here:
Artificial intelligence and the super app - McKinsey

Todays artificial intelligence (AI) systems are only as good as the data theyre trained on. The AI industry is currently taking advantage of large datasets to train AI models and make them more useful. However, as these datasets are becoming limited, researchers are exploring other ways to improve AI algorithms. One such way is quantum computing. It is a new frontier of computer science that will enable better AI algorithms shortly.

Atoms make up our world, and they and their constituents have baffling yet interesting properties. For example, electrons have spin and orbit that can be either up. In addition, they can be in any of the infinite discrete energy levels. These properties determine the quantum states of atoms. At a subatomic level, everything exists as quantum states rather than as traditional logical on or off values. This phenomenon gave rise to quantum computing. It has the potential to change how we see artificial intelligence forever.

Quantum computing is an entirely different way of studying the world around us. It does not just focus on the properties of atoms and molecules. It takes a look at the subatomic properties of atoms that are actually in superposition. That is, they exist in multiple states at the same time. This is one of the principles of quantum mechanics that enable subatomic particles to exist as both particles and waves at the same time.

These principles are strange and counterintuitive. According to them, a computing system cannot only store and process data in binary bits, 0s and 1s. Or in more electronic engineering terms, the state of off and on of an electronic switch. It can also store and process data in superposed states of not on or off but the combination thereof. By harnessing these principles, quantum computers can solve complex problems much faster than traditional computers.

Quantum computers are a variety of different supercomputers based on quantum mechanics. These quantum computers use the laws of quantum mechanics to process information. That means they can find patterns in big data that are almost impossible to find with conventional computers. This way, they are fundamentally different from the computers we use today.

When it comes to artificial intelligence, quantum computing can analyze a wider variety of data. At the same time, they can come to better conclusions than computers today. Conventional computers can only process information as either 1s or 0s. Quantum computers can process information in multiple states known as qubits at once. That enables them to analyze a wider variety of data and come to better conclusions than computers can today.

Artificial intelligence has come a long way in the past few years. It has been able to generate realistic 3D images and videos. In addition, it is beginning to embrace quantum computing. That has given rise to quantum AI. Artificial intelligence now leverages quantum computers. And their full integration will be a technological revolution of the century.

There are several benefits of using quantum AI in creative industries. I have already made it clear it can handle large data sets faster and more efficiently than traditional AI technologies. It can also identify patterns that are difficult for regular computers to spot. Furthermore, it can combine and rearrange existing ideas. Hence it can create new ideas in ways that any human cannot imagine possible.

One of the biggest hurdles for artificial intelligence today is training the machine to do something useful. For example, we might have a model that can correctly identify a dog in a photo. But the model will need to be trained with tens of thousands of images for it to recognize the subtle differences between a beagle, a poodle, and a Great Dane. This process is what AI researchers call training. They use it to teach AI algorithms to make predictions in new situations.

Quantum computing can make this training process faster and more accurate. It will allow AI researchers to use more data than they have ever used before. It can process large amounts of data in 1s and 0s and the combination thereof which will enable quantum computers to come to more accurate conclusions than traditional computers. In other words, AI researchers can use larger datasets to train AI models to be more accurate and better at decision-making.

One of the most exciting predictions for quantum computing in artificial intelligence is the potential to break through language barriers. AI models can currently understand one language the language used to train them. so if we need AI to understand a different language, we shall need to teach it from scratch. However, quantum computing can help AI models break through language barriers. It will allow us to train models in one language and translate them into a different language effortlessly.

That will enable AI to understand and interpret different languages simultaneously. What this will do is create a global AI that can speak multiple languages. Another exciting prediction for the future of AI with quantum computing is the potential to build models with more accurate decision-making skills: Quantum computing will allow using larger datasets to train models. Hence AI will be able to make more accurate decisions that will be especially helpful for financial models, which often have a high rate of inaccuracy because of the limited data used to train them.

Artificial intelligence is already improving the performance of quantum computers. This trend will only continue in the future. The following are some reasons why:

The potential of quantum computing is limitless, but its integration into artificial intelligence will produce a technology that will be rather powerful than anything we have today. The new technology will enable machines to learn and self-evolve. It will make them exponentially better at solving complex problems and developing self-learning algorithms that will drive efficiency in sectors such as finance or healthcare.

Quantum AI systems will be able to process large amounts of information quickly and accurately. That will open up a new world of possibilities for businesses and individuals. They will also be able to solve complex problems that are impossible for even the most advanced conventional computer systems.

Nevertheless, we must remember that these technologies are relatively new; we are still discovering new ways to use quantum computing. Therefore, we must be aware of the latest technology to take advantage of new opportunities as they come along.

The rise of quantum computing will change the way we interact with AI in the future. That means we must stay informed so we can prepare for the changes and make the most of this exciting technology.

Image Credit:

Chisom is a physicist and a tech enthusiast interested in web3, cybersecurity, blockchain application, and quantum information science; he writes about the advancements therein. When he is not writing about tech, he is either crafting absorbing copy for blogs, marketing, and PR, reading his favorite books, or singing at the top of his lungs. He is a shine-through writer with a massive readership.

Original post:
What Quantum Computing Will Mean for the Future Artificial Intelligence - ReadWrite