All posts by Microsoft Cloud Perspectives Blog Team

Curing diseases and delivering effective treatments with the cloud

Researchers trying to cure some of the world’s least-understood diseases, such as Alzheimer’s and Parkinson’s, are discovering new and exciting opportunities in the cloud. With the ability to instantly access vast amounts of computing power, and without the burden of large initial investments or ongoing costs, cloud technology is making it easier for healthcare organizations to study complex disorders and develop innovative new treatments. This is helping lead to an era of more precise and effective medicine.

Although the healthcare industry has used distributed computing networks to tackle large-scale health challenges before—such as the Folding@Home project, which allows individual PC users to contribute unused computing cycles to study how protein misfolding can lead to disease—there are a number of benefits to using a modern cloud computing solution. Cloud-based artificial intelligence (AI) and machine learning (ML) tools, for example, are helping healthcare organizations become more efficient and medical researchers develop better treatments for diseases.

Here are a few ways the cloud is powering healthcare research that’s leading to new cures, while also making the industry more efficient and secure.

Curing Alzheimer’s and Parkinson’s using the cloud

Together, Alzheimer’s and Parkinson’s afflict over 50 million people worldwide. Although the complexity of these disease processes has made finding cures elusive, the cloud’s ability to effortlessly scale resources and instantly tap into significant amounts of computing power is helping researchers gain a better understanding of these diseases faster.

One example: Researchers have built a comprehensive digital model of the human neurological system at comparatively low cost through quicker, on-demand access to high-performance technology stacks. Biotech startup NeuroInitiative uses virtualized GPUs in the cloud to create a general model of the nervous system, a strategy that has proved 40 times faster than using physical GPUs. “The need for lots of GPUs in a pay-as-you-go, easy-to-use model led us to the cloud from day one,” said Andy Lee, co-founder of NeuroInitiative. “We can spin up a 100,000-core cluster in minutes and stop paying for it after experiments finish.”

Using this technology, researchers have been able to simulate a variety of different treatments for neurological diseases. NeuroInitiative alone has identified more than 25 promising drug targets for Parkinson’s. They’ll be ready for human clinical trials in two to three years—half the time it typically takes for preclinical work.

Speeding up treatments with AI and ML

As more healthcare organizations embrace digital technology, they are dealing with increasing volumes of data, which must be quickly processed and analyzed in order to be meaningful. This is just the sort of work cloud-based AI and ML tools are designed to do. Just like on-premises AI and ML systems, these cloud-based tools can help humans handle routine and time-consuming tasks with unmatched speed and accuracy. Plus, these tools don’t require any special knowledge or equipment to set up, making data processing more cost-effective.

However, there’s more to be saved than money. When it comes to the development and distribution of new treatments, fast data analysis can also help combat diseases and save lives. Genomics researchers at the Icahn School of Medicine at Mount Sinai, for example, needed a way to analyze and sequence large sets of genomic data with a limited set of resources. Using cloud-based AI tools from Microsoft Genomics, as well as Azure Data Lake Analytics, they were able to download their data, and then use AI to quickly process and archive it. “The tool can uniformly realign everything and let me do the variant calling for the analysis I want,” said Dr. Robert Klein, head of the Klein Lab at the Icahn School. They are scaling up their research as a result.

Likewise, Intelligent Retinal Imaging Systems (IRIS) wanted to build a platform that could detect diabetic retinopathy, a form of vision loss that can develop rapidly. To help make tests more accessible, they designed a system doctors can use to quickly analyze images of retinas and detect anomalies using ML algorithms. All doctors have to do is send a retinal image to IRIS, which then processes it using Azure Service Bus and Azure Machine Learning Package for Computer Vision. Within 24 hours, doctors receive an enhanced image back with anomalies identified, making it easy for them to give a final diagnosis. “We went from zero to 300,000 patients examined in under five years,” said Jonathan Stevenson, chief strategy and information officer at IRIS. “There is no way we could have done that without Azure.”

Automatically securing healthcare data

Privacy and security have long been concerns in healthcare, and they’ve taken on a much larger significance with the digitization of the industry. Organizations that want to modernize their operations without sacrificing security can use the cloud as an alternative to siloing all their patient records on site.

One of the most effective ways to do this is by choosing a cloud solution that comes with HIPAA and HITRUST compliance. Instead of having to worry about properly storing, accessing, and analyzing sensitive health data, an organization can simply transfer patient information to their cloud, where it will automatically adhere to secure data regulations. Another layer of protection is to use cloud services that emphasize security from the ground up. This includes introducing security elements at every phase, from the initial hardware components to the final transfer of information.

Cloud technology means better treatments are on the way for some of humanity’s most difficult diseases. By giving researchers easy and scalable access to computing resources, cloud technology is helping to reduce the time it takes to test hypotheses, increase the iteration of promising ideas and treatment methods, and develop new cures. It’s also allowing healthcare organizations to modernize without neglecting security. All this is helping to make medicine more effective, accessible, and timely. For doctors, researchers, and patients, the future of healthcare is in the cloud.

To stay up to date on the latest news about Microsoft’s work in the cloud, bookmark this blog and follow us on TwitterFacebook and LinkedIn.

Go to Original Article
Author: Steve Clarke

Feeding the world with AI-driven agriculture innovation

In the 1950s and 1960s, plant biologist Norman Borlaug famously led the “Green Revolution,” developing high-yield grains that helped drive up global food production when paired with innovations in chemical fertilizers, irrigation, and mechanized cultivation. By so doing, Borlaug and his peers helped save a billion people from starvation. However, this new form of farming was not sustainable and created multiple environmental issues.

Today, farmers are using technology to transform production again, driven by the need to feed more with less and to address the impacts of industrial farming on the environment. Currently, nearly half of current food produced, or 2 billion tons a year, ends up as waste, while an estimated 124 million people in 51 countries face food insecurity or worse. In addition, new sources of arable land are limited, fresh water levels are receding, and climate change puts pressure on resources and will lower agricultural production over time. Governments need to solve these issues swiftly, as the world’s population is slated to grow from 7.6 billion to 9.8 billion 2050. Agencies and companies will need to team with growers to drive a 70 percent increase in food production.

The good news is that we’re now in the midst of a second Green Revolution that’s part of the Fourth Industrial Revolution. Here’s how technology innovation, driven by big data, the Internet of Things (IoT), artificial intelligence (AI), and machine learning, will reap a more bountiful harvest.

A vision for AI in agriculture

Farmers are deploying robots, ground-based wireless sensors, and drones to assess growing conditions. They then capitalize on cloud services and edge computing to process the data. By 2050, the typical farm is expected to generate an average of 4.1 million data points every day.

AI and machine learning interpret findings for farmers, helping them continually tweak crop inputs to boost yields. Farmers can use AI to determine the optimal date to sow crops, precisely allocate resources such as water and fertilizer, identify crop diseases for swifter treatment, and detect and destroy weeds. Machine learning makes these activities smarter over time. It can also help farmers forecast the year ahead by using historic production data, long-term weather forecasts, genetically modified seed information, and commodity pricing predictions, among other inputs, to recommend how much seed to sow.

Such precision farming technology augments and extends farmers’ deep knowledge about their land, making production more sustainable. Advanced technology can increase farm productivity by 45 percent while reducing water intake by 35 percent. However, the key is ensuring equitable access: Often the communities that need AI the most lack the physical and technology infrastructure required to support it.

Connecting communities with broadband

Access to high-speed connectivity and reliable power are still challenges in many parts of the world. That’s one reason Microsoft and its partners are bringing affordable broadband to rural communities in countries such as Colombia, India, Kenya, South Africa, and the United States through the Airband Initiative.

When communities are connected, farmers can benefit from AI and machine learning, even if they lack internet access to their individual farms. Microsoft employee Prashant Gupta and his team used advanced analytics and machine learning to create a Personalized Village Advisory Dashboard for 4,000 farmers in 106 villages and a Sowing App for 175 farmers in a district in the southeastern coastal state of Andhra Pradesh in India. Farmers with simple SMS-enabled phones can access Sowing App recommendations, which apply AI to data such as weather and soil conditions to optimize planting times. Farmers who followed the AI-driven advice increased yields by 30 percent over those who adhered to traditional planting schedules.

Using IoT and AI on individual farms

Farmers with connectivity can use IoT to get customized recommendations. The Microsoft FarmBeats program, driven by principal researcher Ranveer Chandra, has developed an end-to-end IoT platform that uses low-cost sensors, drones, and vision/machine learning algorithms to increase the productivity and profitability of farms. FarmBeats is part of Microsoft AI for Earth, a program that provides cloud and AI tools to teams seeking to develop sustainable solutions to global environmental issues.

In the United States, FarmBeats solves the problem of internet connectivity by accessing unused TV white spaces to set up high-bandwidth links between a farmer’s home internet connections and an IoT base station on the farm. Sensors, cameras, and drones connect to this base station, which is both solar- and battery-powered. To avoid unexpected shutdowns due to battery drain, the base station uses weather forecasts to plan its power usage. Similarly, drones leverage an IoT-driven algorithm based on wind patterns to help accelerate and decelerate mid-flight, reducing battery draw.

IoT data processing—for bandwidth-hogging information like drone videos, photos, and sensor feedback—is done by a PC at the farmer’s home. The PC performs local computations and consolidates findings into lower-memory summaries, which can be distributed over bandwidth more easily, while also serving as a backup during network outages.

AI for everyone means more food for the world

Over time, AI will help farmers evolve into agricultural technologists, using data to optimize yields down to individual rows of plants. Farmers without connectivity can get AI benefits right now, with tools as simple as an SMS-enabled phone and the Sowing App. Meanwhile, farmers with Wi-Fi access can use FarmBeats to get a continually AI-customized plan for their lands. With such IoT- and AI-driven solutions, farmers can meet the world’s needs for increased food sustainably—growing production and revenues without depleting precious natural resources.

Be the first to know about new advancements in the Microsoft AI farming initiative. Follow us at FarmBeats.

To stay up to date on the latest news about Microsoft’s work in the cloud, bookmark this blog and follow us on TwitterFacebook, and LinkedIn.

Go to Original Article
Author: Steve Clarke

How the cloud is transforming weather prediction

We’ve come a long way since annual weather forecasts from the Farmer’s Almanac were the basis for how we planned, planted, or invested. Today, there’s a large, growing market for precise, relevant weather information to improve business and government operation. Weather predictions have become a significant strategic advantage, helping ensure uninterrupted service, boosting weather-related sales, improving public safety, and reducing operational risk.

What’s made this possible? At one time, the processing power and vast data storage that were necessary limited weather data crunching to large enterprises that could afford the infrastructure. The cloud has changed all that—dramatically—with vast scalability, lower entry and operating costs, and creation of entirely new business opportunities. Here’s how.

Data in the cloud: sunny outlook for more useful predictions

Tracking constantly changing weather conditions such as temperature, humidity, barometric pressure, wind speeds, and precipitation generates tons of data by the minute. Once it’s gathered, complex calculations are required to make actionable predictions.

The cloud provides efficient storage and processing capabilities that make these tasks much easier. The cloud’s real-time scalability handles the massive and ever-changing data volumes and processing tasks, while using only the resources needed. And since so much information is aggregated, the end product is far more useful. Industry leader AccuWeather’s system automatically compares a customer’s unique sales history with AccuWeather’s historical weather database. A business can upload sales records from a store, a region, or a single SKU, and minutes later get analytics showing how sales of specific products will be affected by incoming rain, snow, or hurricane.

Connectivity creates new business opportunity

The cloud has made meteorology hyper-local. With Internet-of-Things connectivity, it’s now possible to get accurate weather pattern data down to the street level. Fathym uses this capability in a technology package they sell to municipalities and commercial fleets. Vehicles mounted with sensors collect road data that’s synced with weather forecasts. On customized dashboards, users can view current road conditions and set alerts about unsafe driving conditions based on the incoming data. Transportation officials can allocate the right amount of road resources—deicers, snowplows—at the right time.

This powerful cloud/IoT one-two has moved government agencies from reactive to proactive in their winter road management while cutting maintenance and emergency costs. It can also save fleet managers tens of thousands of dollars by rerouting drivers in real time and reducing delays.

Smart machines: making emerging weather patterns more predictable

AccuWeather is using cloud-based machine learning to discover new patterns in data, deliver lifesaving weather data in new and innovative ways, and provide up-to-date predictions that drive profits.

A leader in high-performance computing, the company is aggressively adopting the cloud to improve forecasts, scale its services, and innovate. “The impetus for moving to the cloud was better scalability and resilience, which we need for delivering timely information such as tornado warnings to our customers,” says chief technology officer Chris Patti.

AccuWeather recently launched its D3 Express cloud-based analytics product that quantifies and rapidly notifies everyone—from farmers to holiday retailers—of disruptive weather events. Like never before, these clients can make decisions that maximize returns on weather-driven opportunities and minimize losses. With cloud-enabled, machine-learning services, AccuWeather can scale such advanced modeling across its entire customer base.

In addition, the company has migrated its APIs to the cloud, simplifying scalability, ensuring high-quality, reliable access to valuable weather data for customers—and reducing the company’s operational overhead to boot.

The sky’s the limit

The cloud has lowered the cost of entry for organizations that want to capitalize on real-time weather science. There’s less hardware to own. Off-the-shelf, advanced analytics are available from pre-built services, along with plug-and-play APIs. All this is making weather data more accessible to more users.

At a time when weather patterns are veering into new territory, the forecast calls for clear skies for a cloud-first approach that’s bringing more predictability on the weather. Learn more about how to get actionable insights from cloud-based data.

To stay up to date on the latest news about Microsoft’s work in the cloud, bookmark this blog and follow us on Twitter, Facebook, and LinkedIn.

The new business imperative: A unified cloud security strategy

As more businesses begin to explore the benefits of moving on-premises data and applications to the cloud, they’re having to rethink their traditional approaches to data security. Not only are cybercriminals developing more sophisticated attacks, but the number of employees and users who can access, edit, and share data has increased the risk of breaches. In fact, Gartner indicates* that “through 2022, 95 percent of cloud security incidents failures will be the customer’s fault. CIOs can combat this by implementing and enforcing policies on cloud ownership, responsibility and risk acceptance. They should also be sure to follow a life cycle approach to cloud governance and put in place central management and monitoring planes to cover the inherent complexity of multicloud use.”

Instead of relying on a patchwork of third-party security solutions that don’t always speak to each other, potentially leaving systems vulnerable to attack, companies are now adopting a unified, end-to-end cloud security defense. This typically involves choosing a cloud provider that can integrate security controls right into existing corporate systems and processes. When these controls span the entire IT infrastructure, they make it easier to protect data and maintain user trust by offering increased compatibility, better performance, and more flexibility.

Protection that’s always compatible

A holistic, cloud-supported threat warning and detection system can be designed to work seamlessly across every asset of an IT environment. For instance, built-in security management solutions can give IT teams the ability to constantly monitor the entire system from a centralized location, rather than manually evaluating different machines. This allows them to sense threats early, provide identity monitoring, and more—all without any compatibility issues.

Container shipping company Mediterranean Shipping Company (MSC) has gone this route. As in many businesses, MSC’s IT environment is spread across a variety of locations, networks, and technologies, such as container ships, trucking networks, and offices. Its previous security strategy employed a mixture of third-party solutions that often ran into compatibility issues between different components, giving attackers a large surface area to probe. This made MSC vulnerable to threats such as fileless attacks, phishing, and ransomware. However, after transitioning to a unified cloud security solution, it has been able to guard against attacks using protection that integrates effortlessly into its existing environment.

Reliable performance, more efficiently

The more complex an IT environment gets, the more time employees spend testing, maintaining, and repairing third-party security solutions. A unified cloud security approach improves performance by not only providing a consistent, layered defense strategy, but by also automating it across the entire IT infrastructure. At MSC, software and security updates are now done automatically and deployed without delay across the cloud. Information about possible threats and breaches can quickly be shared across devices and identities, speeding up response and recovery times so that employees can focus on other issues.

Security with flexibility to grow

Scalability is another factor driving adoption. A cloud environment can easily scale to accommodate spikes in traffic, additional users, or data-intensive applications. A patchwork of third-party security solutions tends not to be so nimble. At MSC, security controls are integrated into multiple levels of the existing IT infrastructure—from the operating system to the application layer—and can be dynamically sized to meet new business needs. For example, continuous compliance controls can be established to monitor regulatory activities and detect vulnerabilities as they grow.

A unified security approach: becoming the standard

The best security solutions perform quietly in the background, protecting users without them noticing. Unified cloud security does this while also reducing the resources required to keep things running smoothly. “Once you have true defense in depth, there’s less chance of having to single out a user and impact their productivity because you have to reimage an infected machine,” said Aaron Shvarts, chief security officer at MSC Technology North America.

After moving its workloads to Azure and upgrading its previous third-party security solutions to the native protection of Windows Defender, MSC now has a defense strategy that suits the complexity of its business. Learn more about Azure security solutions and how Microsoft can help you implement unified security across your cloud.

To stay up to date on the latest news about Microsoft’s work in the cloud, bookmark this blog and follow us on Twitter, Facebook, and LinkedIn.

*Gartner, Smarter with Gartner, Is the Cloud Secure?, 27 March 2018,

In 2018, a better, faster, more accessible cloud emerges

Here’s what’s new in the Microsoft Cloud: Microsoft is making it easier for developers to build great apps that take advantage of the latest analytics capabilities with free developer tools and languages, best-practice guidance, price reductions, and new features.

Better decisions through better analytics

Knowing how users interact with your apps is a critical first step in managing product strategy and development pipeline. Using robust analytics, you can get the immediate feedback you need to determine how to engage users and make better decisions to improve your apps. With Visual Studio App Center, you can access App Center Analytics completely free. Now you can use this tool with Azure Application Insights to improve your business. Get started today.

New tools speed app development using time series data

Integrating IoT with other real-time applications can be a complex challenge. With Time Series Insights (TSI), developers can build applications that give valuable insights to customers, take fine-grain control over time series data, and easily plug TSI into a broader workflow or technology stack. To help developers get started and shorten development cycles, Microsoft has released new Azure Time Series Insights developer tools. With these tools, developers can more easily embed TSI’s platform into apps to power charts and graphs, compare data from different points in time, and dynamically explore data trends and correlations.

Faster feedback drives better apps

Good intuition is important, but without user input and insights you are playing a potentially costly guessing game. Gathering feedback fast from beta users who are invested in your product’s success lets you learn and adapt quickly before getting too deep into code that’s expensive to correct later. Using this step-by-step guide from one of our Visual Studio App Center customers, you will learn how to swiftly gather quantitative and qualitative user feedback to build apps your customers love, anticipate and correct problems, and ultimately win customers’ loyalty.

Empowering data scientists with R updates

R, an open-source statistical programming language, empowers data scientists to drive insightful analytics, statistics, and visualizations for mapping social and marketing trends, developing scientific and financial models, and anticipating consumer behavior. Recently we’ve released Microsoft R Open 3.4.3, the latest version of Microsoft’s enhanced distribution of R. This free download includes the latest R language engine, compatibility, and additional capabilities for performance, reproducibility, and platform support.

New open-source analytics capabilities at a lower cost

Microsoft recently announced significant price reductions, along with new abilities for Azure HDInsight, the open-source analytics cloud service that developers can implement in a wide range of mission-critical applications, including machine learning, IoT, and more. This includes capabilities like Apache Kafka on Azure HDInsight and Azure Log Analytics integration, previews for Enterprise Security Package for Azure HDInsight, and integration with Power BI direct query.

We are constantly creating new tools and features that reduce time-to-market and allow developers to do their best work. To stay up to date on Microsoft’s work in the cloud, visit