Developing for all 1 billion Windows 10 devices and beyond – Windows Developer Blog

This year, Microsoft Build 2020 is a digital-only event that we all get to experience from the comfort of our homes. We hope you enjoy learning about the new features and technologies that matter most to you. Today, I will have the privilege of sharing how developers can build apps for modern work using Microsoft 365 and Windows platforms. I will focus on 4 key areas of improvements to the Windows platform:
Unifying app development across the billion Windows 10 devices for all your current and future apps;
Leaning into the cloud and enabling new scenarios for your Windows apps;
Creating new opportunities for you to build connected apps using Microsoft 365 integration in the Windows experience; and
Making Windows great for developer productivity.

Today we will unveil Project Reunion: our vision for unifying and evolving the Windows developer platform to make it easier to build great apps that work across all the Windows 10 versions and devices people use.
For the past couple of years, we have been breaking down the barrier between Win32 (also called the Windows API) and Universal Windows Platform (UWP) APIs. Project Reunion expands this effort to make it easier to build a great Windows app. It will unify access to existing Win32 and UWP APIs and make them available decoupled from the OS, via tools like NuGet. This will provide a common platform for new apps. Plus, it will help you update and modernize your existing apps with the latest functionality, whether they’re C++, .NET (including WPF, Windows Forms, and UWP) or React Native. As we decouple existing APIs and add new APIs, we are also doing the work to polyfill, as needed, so the APIs work down-level across supported versions of Windows.

One of the first components in the Project Reunion journey is WinUI 3 Preview 1, the high performant, Fluent-optimized native UI framework for Windows. With WinUI developers can build great user experiences that adapt and scale across devices, whether they are starting a new project, or modernizing an existing app.

Image 1: Engaging UI powered by WinUI in Alarms & Clock app
We also know there are times when you want to integrate web content in your native app so you can share code across platforms and with the browser. Today, we are expanding WebView2 (another component in the Project Reunion journey) with a new .NET Preview. Now, any Windows app can embed web content with the power of Microsoft Edge and Chromium. WebView2 provides full web functionality across the spectrum of Windows apps, and it’s decoupled from the OS, so you are no longer locked to a particular version of Windows.

Image 2: Showing an example of a PDF inline using WebView2
We invite you to learn and engage with us at this early stage in the Project Reunion GitHub repo, where we’ll be sharing our progress and listening to your feedback as we implement this vision. You can also learn more about WinUI and WebView2 today.

As organizations shift to enable remote work, Windows Virtual Desktop, built on Azure, has provided the ability to provision and scale virtual desktops and apps faster than was previously possible. It enables organizations to serve your existing app to a growing set of devices that users can access with the Windows Virtual Desktop clients for Windows, MacOS/iOS, and Android.
Because scalability is so important, we introduced a feature called MSIX App Attach. The MSIX App Attach Preview will optimize people’s experiences by untangling the OS image that organizations deploy in the cloud from the apps that people need to access. This means that once you adopt MSIX for Windows desktop, the same investment will soon bring even more benefits when running your app in Windows Virtual Desktop on Azure.

We also know you are looking for more ways to build connect apps, and with Microsoft 365 integrations like Microsoft Search and the Microsoft Graph you have even more opportunities. We all use so many tools and apps and it can be cumbersome to find what we need. But Microsoft Search and the Microsoft Graph can draw unique connections between your people, files, and tools so that you can find what you are looking for. These are just two integrations that allow you and your users to be more efficient, but you can do so much more with the Microsoft 365 platform. Additionally, developers can start using our Graph Connectors that are in preview now, and Microsoft Search will be coming to Windows later this year.

Image 3: Look for a coworker and see pertinent information and shared files and apps

We know it is important for you to stay productive and we appreciate your feedback on how we can make that happen. With advancements to Windows Terminal and Windows Subsystem for Linux you have modern, fast, and powerful tools.
Now available for enterprise use, Windows Terminal 1.0 provides you with the ability to run any command line executable, including WSL distros and Azure Cloud Shell, inside multiple tabs, and panes. You can also use Unicode and UTF-8 characters, have a GPU accelerated text rendering engine, and custom themes, styles, and configurations. The Windows Terminal is available on the Microsoft Store or manually from the Terminal GitHub repo.
Improvements to Windows Subsystem for Linux (WSL) have centered around enabling hardware acceleration, running a Linux GUI app directly, and making it easier to start using Linux apps on Windows. Here are a few details:
Added support for graphics processing unit (GPU) compute workflows allows Linux tools to leverage GPUs to enable hardware acceleration for many development scenarios, such as parallel computation and training machine learning (ML) and artificial intelligence (AI) models.
Support for Linux graphical user interface (GUI) apps will enable you to open a WSL instance and run a Linux GUI app directly without the need for a third-party X server. This will help you to run your favorite apps in a Linux environment such as an integrated development environment (IDE).
WSL will soon support a simplified install experience by running the command ‘wsl.exe – install,’ which will make it easier than ever to start using Linux apps on Windows.
Additionally, preview tools and utilities, like the Windows Package Manager and Microsoft PowerToys, provide you with paths to streamline your Windows experience for even greater development productivity.
You asked for an easier way to setup your development environment and now with the Windows Package Manager Preview you have a command line interface enabling you to install your favorite tools quickly and easily. The repository of packages is open source, and we can’t wait for you to contribute and help us with the next level of improvements.
With Microsoft PowerToys (0.18) you can customize the Windows 10 shell for your personal workflows. Today’s updates add two new utilities: Keyboard Remapper and PowerToys Run. You can remap key to key and shortcut to shortcut using Keyboard Remapper. And, PowerToys Run, an app launcher utility gets you to your programs faster than before – hit alt-space and just start typing.
We look forward to working with you in the open to make progress on each of these efforts, so we can help you build productive and delightful experiences. I can’t wait to see what we can build together with WinUI, WebView2, Terminal, Windows Package Manager Preview, Project Reunion and more. If you didn’t get enough at Build, you can find additional deep-dive content on topics like WSL, Terminal, React Native for Windows, WebView2, Windows AI, and much more at Microsoft365.

Introducing WinUI 3 Preview 1 – Windows Developer Blog

Today at Microsoft Build 2020 the WinUI team is excited to announce the latest version of WinUI 3 available for public preview: WinUI 3 Preview 1. This preview expands the world of WinUI to an even wider audience of developers – the previously-released WinUI 3 Alpha was only available to UWP app developers, but Preview 1 now enables WinUI usage in Desktop (Win32) apps as well!

WinUI is Microsoft’s most advanced user interface technology for building Windows apps. There are two versions of WinUI in active development: WinUI 2 and WinUI 3. WinUI 2 is a library of Fluent-based UI controls & styles for UWP XAML apps; it was first shipped in Oct 2018 and its latest release is v2.4 which came out recently on May 8th.
WinUI 3 is a currently-in-development, dramatic expansion of this library into a full-fledged, end-to-end, standalone UI framework. This UI framework continues the tradition of WinUI and UWP XAML, providing the very latest graphical capabilities and Fluent Design styling that embraces today’s modern devices, hardware, and inputs. Technically, WinUI 3 decouples the XAML, Composition, and Input layers of Windows 10, and ships them independently via NuGet for any app targeting Windows 10 1803 and above. It can be used in both C++ and .NET-based apps.
WinUI 3 Preview 1 is the first pre-release of WinUI 3 that can be used in both UWP and Desktop apps, allowing every developer a chance to tap into the very latest UI technologies from Microsoft for building Windows apps. You can learn more about WinUI here.

In addition to the aforementioned ability to use WinUI from Desktop apps, Preview 1 will also support .NET 5 when run in this Desktop context. Preview 1 is also the first preview of WinUI 3 to include the SwapChainPanel control – this is key for developers working with DirectX for providing custom graphical rendering as a part of their user experience.
In February, the WinUI team released an Alpha of WinUI 3 with an all-new Chromium-based WebView2 control. This same control carries through to Preview 1 and has been updated with several improvements. Preview 1 also offers several other advancements compared to the February Alpha:
Support for the .NET versions of several interfaces when using .NET 5. In Preview1, .NET 5 is used with Desktop apps. The interfaces (and a class) are: INotifyPropertyChanged, INotifyCollectionChanged, ObervableCollection<T>, and ICommand.
New, recently released controls ported forward from WinUI v2.4.
Several important bug fixes & stability improvements.
A full articulation of the capabilities and limitations of Preview 1 can be found in the WinUI 3 Preview 1 Release Notes. Of special note: Preview 1 should not be used in production apps, as it does not guarantee forward compatibility and may break in future updates we make to Windows 10.

At Build 2020 we’ll share stories of customers and ecosystem partners we’re working with to improve & develop WinUI 3. Examples of these partnerships that will be referenced during the conference include:
Magix, a developer of audio & video solutions for creators, who’s showcasing how they’ve harnessed the power of WinUI to modernize some of the UI components in their Win32-based VEGAS Pro app
Esri, market leader in geographic information system (GIS) software, using ArcGIS Runtime, .NET and SwapChainPanel with WinUI 3 Preview 1 to render gigabytes of geospatial data in rich, high performance mapping components
Telerik, a leading vendor of high quality .NET and JavaScript components for building modern, high performance, feature rich web, mobile and desktop apps
Uno platform, an open source platform for building apps that span across mobile endpoints, Windows, macOS, and the Web
These are just some of the great companies exploring and engaging with us as we develop WinUI 3, and in the coming months you’ll learn more about these and other engagements as we prepare for the eventual release of WinUI 3.0 for general availability.

To start with WinUI 3.0 Preview 1, head over to our download and release notes page. You can also find more information and technical setup details in our documentation on creating a Desktop app with WinUI.

Everything new from Microsoft Edge at Build 2020 – Microsoft Edge Blog

Microsoft Build 2020 can be defined by a set of ‘firsts’—it’s the first Build of the new decade, the first Build that’s all-digital for 24 hours a day, and the first Build since the new Microsoft Edge launched! We’re learning a lot as we deliver Build and Microsoft Edge in new ways.
Last year at Build, we unveiled what was coming with the new Microsoft Edge. Since then we haven’t stopped innovating to build a browser that’s right for you, with world-class performance and more security, more productivity, and more value as you browse.
We’ve introduced tools to give you more control over your privacy with Tracking prevention. We’ve added new ways to be more productive with Collections. And with Give Mode, we’ve turned searching with Bing into an easy way to do something good.
As mentioned in January, we plan to upgrade all Windows 10 devices (excluding enterprise and education) to the new Microsoft Edge. Users will be able to experience all the features above when it’s delivered via a measured roll-out that you’ll see ramping up over the course of the next few weeks. If you can’t wait and want to try it now, you can download the new Microsoft Edge here.
We believe in a web that just works for everyone. As we work to deliver a great browsing experience, we remain committed to listening to developers and sharing improvements back to the open source community that makes Microsoft Edge possible. To date, we’ve made over 3,000 commits back to the Chromium open source project that make the web more enjoyable, capable, and accessible for everyone.
This year at Build, we’re sharing ongoing innovation with improved tooling and reach for web developers, new consumer integrations, and new customization options for businesses.

What’s new for web developers
A healthy web community is made possible by a vibrant developer community. Over the last few months, we’ve worked to support that community with improvements like DevTools localization in 10 new languages, which has been adopted by many of you as you develop for Microsoft Edge. Now, we’re excited to highlight new tools that empower you, the developers who make the web possible, while expanding your canvas so you can reach your customers in more ways than ever before.
WebView2 preview expands to include .NET and UWP (WinUI) development

Last year, we pushed WebView2 forward with a preview for Win32 development. WebView2 lowers the barrier for developers to maximize code reuse across platforms with a consistent web platform to host web content in their apps. We’d like to thank everyone that has engaged with us so far throughout the preview—the contributions and feedback we’ve received drive our feature roadmap and quality.
Today, we’re expanding the preview with new options for .NET and UWP (WinUI 3.0) development, enabling you to embed a Chromium-based Edge WebView in WinForms, WPF, and UWP (WinUI 3.0) applications. Check out our documentation and Getting Started guide, or simply open Visual Studio and download the WebView2 package to get started.
Improvements to the Microsoft Edge Add-ons site make searching and finding extensions easier
For many users, extensions are a key component to the web experience. We’re committed to not only making it easy for developers to bring their Chromium-based extensions into our store, but also to make it simple for users to find them. We will be making a significant update to the Microsoft Edge Add-ons site to make it more visually appealing—new categories, new search capabilities, and a new layout will get those extensions in front of customers. Expect to see this roll out starting this month!
Making PWAs feel more at home on Windows 10
Progressive Web Apps (PWAs) continue to gain momentum as a great way to deliver content in an app-like experience across platforms, powered by web standards. With the most recent versions of Microsoft Edge and Windows 10, we’re making PWAs feel even more natural and familiar alongside other Windows apps.
Today, when you install PWAs on Windows 10, they look and feel more at home on your desktop OS in important ways, launching in a standalone window and appearing in places like Start and the Taskbar for easier access. We’re working on updates to make this integration even smoother—you’ll be able to manage them from settings, use them to share (and receive shared content), and more. Check out an early preview of this Windows integration by using Windows Insider preview builds and enabling the Web Apps Identity Proxy flag in the Microsoft Edge Canary preview build.
Try out experimental web platform features with Origin Trials
We always aim to move the web forward to meet the needs of web developers. Our Origin Trials program enables developers to test drive experimental features on their websites for a set length of time. Prototypes that we haven’t enabled for the general web yet will work on your site for a selection of your visitors in Microsoft Edge, enabling you to gather and provide early feedback which can influence the final API.
Learn more and register for an Origin Trial at the new Origin Trial Developer Console today!
New features for every user
This week, we’re also introducing a few exciting new features to help you be more productive, and a new integration to help find and organize the content you want.
Collections in Microsoft Edge is getting smarter with a new Pinterest integration
Collections helps you organize, save, and share your online life by allowing you to create groups of content from around the web. Links, images, text, and notes can be saved together across multiple sites, pages, and browser sessions to help you stay organized and inspired.
Now, Collections in Microsoft Edge is collaborating with Pinterest to help you discover relevant content for your research or project. People use Collections to capture their favorite design ideas, recipes, home improvement links, and even to research their next big purchase across multiple sites. This new integration with Pinterest will help you find ideas to spark inspiration, save time, and be more productive by suggesting content related to what you have already collected. When you enable the feature, you will see Pinterest suggestions at the bottom of your collection. Clicking on a suggestion will open a board of similar, trending Pins so you can quickly find and add ideas relevant to you.
This collaboration also allows you to export your collection to Pinterest. Any saved webpages or images will then show up in a new board in your Pinterest account.

Collections will also be rolling out the ability to send to OneNote, in addition to the options to send to Excel and Word which are available today. Collections integration with Pinterest, and Send to OneNote, will appear in Insider channels within the next month. Join our Edge insider program to be one of the first people to try it.
Sidebar search provides a faster, more contextual way to look things up

Today you have the option to look up words or phrases by searching in a new tab. This can make you lose your train of thought instead of helping you get more out of what you’re reading. Sidebar search aims to improve this experience by giving you the option to see results in a pane on the side of the page. Simply highlight a word or phrase, right click, and select “search in sidebar” from the context menu. And if you’re at work and signed in with your Azure Active Directory account, you’ll even see company results! Reading a corporate strategy paper with unlinked references? Use sidebar search to help find the answer. When you’re done, you can close out of the pane or keep searching in it to look up additional information.
Expect to see this show up in Insider preview channels in the coming weeks. Become an Edge Insider to be one of the first to try it.
New features for information workers and IT Pros
Microsoft Edge is the browser for business, and today we’re announcing new features and updates to make the lives of IT managers and information workers a little bit easier. Want to try these out in your organization? Download offline installers here.
New syncing and customization options for IT professionals

If you use multiple devices while working from home, you know how important sync is. Installed extensions now sync in Microsoft Edge and a new policy even allows IT professionals to precisely manage which types of data sync for their users. The flexibility to sync individual data types means IT professionals can fit syncing to their workplace needs. For some workplaces, syncing passwords might not be allowed—now IT can manage this.
Coming soon, sync will also extend to customers whose environments are still on-premises. The transition to the cloud takes time, and your company should be able to get the most out of Microsoft Edge during that transition.
Data security gets a boost with Windows Information Protection
As many people work from home, data protection is top of mind, especially for IT professionals. Microsoft Edge now supports Windows Information Protection for Windows 10 customers which clearly separates personal and corporate data, adds extra protection for line-of-business apps, and provides audit reporting for compliance. This has been a top ask by many customers, and we’re excited to bring it to the new Microsoft Edge.
Easily move between work and personal profiles

For information workers, we’ve made improvements to help keep them in their flow throughout the workday—this is especially true for those working at home, where work and personal boundaries can easily blend. Microsoft Edge now allows users to set a default profile for any link they open, creating a consistent experience even if they’re switching between profiles throughout the day.
Microsoft Edge helps you manage your profiles even more with a feature called Automatic Profile Switching. Previously, if you’ve been using your personal account and then try to access a work link, you’d be forced to sign-in again because your personal account doesn’t have your work credentials. With the new Automatic Profile Switching feature, Microsoft Edge will detect that the link you’re trying to open needs work credentials, and then switch you to your work profile to open it. Profile switching has never been smoother.
Microsoft Search in Bing puts your work results all in one place
Bing is bringing the life-changing capability of search to the workplace, and we have an exciting update as part of this year’s Build. A comprehensive Work page will now be part of Bing’s search results pages for all Microsoft 365 customers. When signed-in to Bing with your work credentials, this page will appear right next to other familiar pages such as Images, Shopping, and News. This new results page gives you the option to view work-specific results such as files, people, internal websites, and more. If your organization uses Microsoft 365 Apps for enterprise, just sign into Bing with your work credentials and try it today. Learn more on our new website!
We hope you enjoy this new 48-hour, all-digital format as we meet you where you are around the world—we truly can’t thank you enough for joining us online for Microsoft Build 2020.
Keep innovating, keep building, and keep expecting more from the web.

When considering responsible AI, begin with the who

At the 2005 Conference on Neural Information Processing Systems, researcher Hanna Wallach found herself in a unique position—sharing a hotel room with another woman. Actually, three other women to be exact. In the previous years she had attended, that had never been an option because she didn’t really know any other women in machine learning. The group was amazed that there were four of them, among a handful of other women, in attendance. In that moment, it became clear what needed to be done. The next year, Wallach and two other women in the group, Jennifer Wortman Vaughan and Lisa Wainer, founded the Women in Machine Learning (WiML) Workshop. The one-day technical event, which is celebrating its 15th year, provides a forum for women to present their work and seek out professional advice and mentorship opportunities. Additionally, the workshop aims to elevate the contributions of female ML researchers and encourage other women to enter the field. In its first year, the workshop brought together 100 attendees; today, it draws around a thousand.

In creating WiML, the women had tapped into something greater than connecting female ML researchers; they asked whether their machine learning community was behaving fairly in its inclusion and support of women. Wallach and Wortman Vaughan are now colleagues at Microsoft Research, and they’re channeling the same awareness and critical eye to the larger AI picture: Are the systems we’re developing and deploying behaving fairly, and are we properly supporting the people building and using them?

Senior Principal Researchers Jennifer Wortman Vaughan (left) and Hanna Wallach (right), co-founders of the Women in Machine Learning Workshop, bring a people-first approach to their work in responsible AI. The two have co-authored upward of 10 papers together on the topic, and they each co-chair an AI, Ethics, and Effects in Engineering and Research (Aether) working group at Microsoft.

Wallach and Wortman Vaughan each co-chair an AI, Ethics, and Effects in Engineering and Research (Aether) working group—Wallach’s group is focused on fairness, Wortman Vaughan’s on interpretability. In those roles, they help inform Microsoft’s approach to responsible AI, which includes helping developers adopt responsible AI practices with services like Azure Machine Learning. Wallach and Wortman Vaughan have co-authored upward of 10 papers together around the topic of responsible AI. Their two most recent publications in the space address the AI challenges of fairness and interpretability through the lens of one particular group of people involved in the life cycle of AI systems: those developing them.

“It’s common to think of machine learning as a fully automated process,” says Wortman Vaughan. “But people are involved behind the scenes at every step, making decisions about which data to use, what to optimize for, even which problems to solve in the first place, and each of these decisions has the potential to impact lives. How do we empower the people involved in creating machine learning systems to make the best choices?”

Their findings are presented in “Co-Designing Checklists to Understand Organizational Challenges and Opportunities around Fairness in AI” and “Interpreting Interpretability: Understanding Data Scientists’ Use of Interpretability Tools for Machine Learning.” The publications received ACM CHI Conference on Human Factors in Computing Systems (CHI 2020) best paper recognition and honorable mention, respectively.

A framework for thinking about and prioritizing fairness

When Wallach took the lead on the Aether Fairness working group, she found herself getting the same question from industry colleagues, researchers in academia, and people in the nonprofit sector: Why don’t you just build a software tool that can be integrated into systems to identify issues of unfairness? Press a button, make systems fair. Some people asked in jest; others more seriously. Given the subjective and sociotechnical nature of fairness, there couldn’t be a single tool to address every challenge, and she’d say as much. Underlying the question, though, was a very real truth: Practitioners needed help. During a two-hour car ride while on vacation, Wallach had an aha moment listening to a Hidden Brain podcast episode about checklists. What practitioners wanted was a framework to help them think about and prioritize fairness.

“I’m getting this question primarily from people who work in the technology industry; the main way they know how to ask for structure is to ask for software,” she recalls thinking of the requests for a one-size-fits-all fairness tool. “But what they actually want is a framework.”

Wallach, Wortman Vaughan, Postdoctoral Researcher Luke Stark, and PhD candidate Michael A. Madaio, an intern at the time of the work, set out to determine if a checklist could work in this space, what should be on it, and what kind of support teams wanted in adopting one. The result is a comprehensive and customizable checklist that accounts for the real-life workflows of practitioners, with guidelines and discussion points for six stages of AI development and deployment: envision, define, prototype, build, launch, and evolve.

During the first of two sets of workshops, researchers presented participants with an initial AI fairness checklist culled from existing lists, literature, and knowledge of fairness challenges faced by practitioners. Participants were asked to give item-level feedback using sticky notes and colored dots to indicate edits and difficulty level of accomplishing list items, respectively. The researchers used the input to revise the checklist.

Co-designing is key

AI ethics checklists and principles aren’t new, but in their research, Wallach, Wortman Vaughan, and their team found current guidelines are challenging to execute. Many are too broad, oversimplify complex issues with yes/no–style items, and—most importantly—often appear not to have included practitioners in their design. Which is why co-designing the checklist with people currently on the ground developing AI systems formed the basis of the group’s work.

The researchers conducted semi-structured interviews exploring practitioners’ current approaches to addressing fairness issues and their vision of the ideal checklist. Separately, Wallach, Wortman Vaughan, and others in the Aether Fairness working group had built out a starter checklist culled from existing lists and literature, as well as their own knowledge of fairness challenges faced by practitioners. The researchers presented this initial checklist during two sets of workshops, revising the list after each based on participant input regarding the specific items included. Additionally, the researchers gathered information on anticipated obstacles and best-case scenarios for incorporating such a checklist into workflows, using the feedback, along with that from the semi-structured interviews, to finalize the list. When all was said and done, 48 practitioners from 12 tech companies had contributed to the design of the checklist.

During the process, researchers found that fairness efforts were often led by passionate individuals who felt they were on their own to balance “doing the right thing” with production goals. Participants expressed hope that having an appropriate checklist could empower individuals, support a proactive approach to AI ethics, and help foster a top-down strategy for managing fairness concerns across their companies.

A conversation starter

While offering step-by-step guidance, the checklist is not about rote compliance, says Wallach, and intentionally omits thresholds, specific criteria, and other measures that might encourage teams to blindly check boxes without deeper engagement. Instead, the items in each stage of the checklist are designed to facilitate important conversations, providing an opportunity to express and explore concerns, evaluate systems, and adjust them accordingly at natural points in the workflow. The checklist is a “thought infrastructure”—as Wallach calls it—that can be customized to meet the specific and varying needs of different teams and circumstances.

During their co-design workshops, researchers used a series of storyboards based on participant feedback to further understand the challenges and opportunities involved in incorporating AI fairness checklists into workflows.

And just as the researchers don’t foresee a single tool solving all fairness challenges, they don’t view the checklist as a solo solution. The checklist is meant to be used alongside other methods and resources, they say, including software tools like Fairlearn, the current release of which is being demoed this week at the developer event Microsoft Build. Fairlearn is an open-source Python package that includes a dashboard and algorithms to support practitioners in assessing and mitigating unfairness in two specific scenarios: disparities in the allocation of opportunities, resources, and information offered by their AI systems and disparities in system performance. Before Fairlearn can help with such disparities, though, practitioners have to identify the groups of people they expect to be impacted by their specific system.

The hope is the checklist—with such guidance as “solicit input on system vision and potential fairness-related harms from diverse perspectives”—will aid practitioners in making such determinations and encourage other important conversations.

“We can’t tell you exactly who might be harmed by your particular system and in what way,” says Wallach. “But we definitely know that if you didn’t have a conversation about this as a team and really investigate this, you’re definitely doing it wrong.”

Tackling the challenges of interpreting interpretability

As with fairness, there are no easy answers—and just as many complex questions—when it comes to interpretability.

Wortman Vaughan recalls attending a panel discussion on AI and society in 2016 during which one of the panelists described a future in which AI systems were so advanced that they would remove uncertainty from decision-making. She was confounded and angered by what she perceived as a misleading and irresponsible statement. The uncertainty inherent in the world is baked into any AI systems we build, whether it’s explicit or not, she thought. The panelist’s comment weighed on her mind and was magnified further by current events at the time. The idea of “democratizing AI” was gaining steam, and models were forecasting a Hillary Rodham Clinton presidency, an output many were treating as a done deal. She wondered to the point of obsession, how well do people really understand the predictions coming out of AI systems? A dive into the literature on the ML community’s efforts to make machine learning interpretable was far from reassuring.

“I got really hung up on the fact that people were designing these methods without stopping to define exactly what they mean by interpretability or intelligibility, basically proposing solutions without first defining the problem they were trying to solve,” says Wortman Vaughan.

That definition rests largely on who’s doing the interpreting. To illustrate, Wallach provides the example of a machine learning model that determines loan eligibility: Details regarding the model’s mathematical equations would go a long way in helping an ML researcher understand how the model arrives at its decisions or if it has any bugs. Those same details mean little to nothing, though, to applicants whose goal is to understand why they were denied a loan and what changes they need to make to position themselves for approval.

In their work, Wallach and Wortman Vaughan have argued for a more expansive view of interpretability, one that recognizes that the concept “means different things to different people depending on who they are and what they’re trying to do,” says Wallach.

As ML models continue to be deployed in the financial sector and other critical domains like healthcare and the justice system—where they can significantly affect people’s livelihood and well-being—claiming ignorance of how an AI system works is not an option. While the ML community has responded to this increasing need for techniques that help show how AI systems function, there’s a severe lack of information on the effectiveness of these tools—and there’s a reason for that.

“User studies of interpretability are notoriously challenging to get right,” explains Wortman Vaughan. “Doing these studies is a research agenda of its own.”

Not only does designing such a study entail qualitative and quantitative methods, but it also requires an interdisciplinary mix of expertise in machine learning, including the mathematics underlying ML models, and human–computer interaction (HCI), as well as knowledge of both the academic literature and routine data science practices.

The enormity of the undertaking is reflected in the makeup of the team that came together for the “Interpreting Interpretability” paper. Wallach, Wortman Vaughan, and Senior Principal Researcher Rich Caruana have extensive ML experience; PhD student Harmanpreet Kaur, an intern at the time of the work, has a research focus in HCI; and Harsha Nori and Samuel Jenkins are data scientists who have practical experience building and using interpretability tools. Together, they investigated whether current tools for increasing the interpretability of models actually result in more understandable systems for the data scientists and developers using them.

Three visualization types for model evaluation are output by the popular and publicly available InterpretML implementation of GAMs (top) and the implementation of SHAP in the SHAP Python package (bottom), respectively. Left column: global explanations. Middle column: component (GAMs) or dependence plot (SHAP). Right column: local explanations.

Tools in practice

The study focuses on two popular and publicly available tools, each representative of one of two techniques dominating the space: the InterpretML implementation of GAMs, which uses a “glassbox model” approach, by which models are designed to be simple enough to understand, and the implementation of SHAP in the SHAP Python package, which uses a post-hoc explanation approach for complex models. Each tool outputs three visualization types for model evaluation.

Through pilot interviews with practitioners, the researchers identified six routine challenges that data scientists face in their day-to-day work. The researchers then set up an interview study in which they placed data scientists in context with data, a model, and one of the two tools, assigned randomly. They examined how well 11 practitioners were able to use the interpretability tool to uncover and address the routine challenges.

The researchers found participants lacked an overall understanding of the tools, particularly in reading and drawing conclusions from the visualizations, which contained importance scores and other values that weren’t explicitly explained, causing confusion. Despite this, the researchers observed, participants were inclined to trust the tools. Some came to rely on the visualizations to justify questionable outputs—the existence of the visualizations offering enough proof of the tools’ credibility—as opposed to using them to scrutinize model performance. The tools’ public availability and widespread use also contributed to participants’ confidence in the tools, with one participant pointing to its availability as an indication that it “must be doing something right.”

Following the interview study, the researchers surveyed nearly 200 practitioners, who were asked to participate in an adjusted version of the interview study task. The purpose was to scale up the findings and gain a sense of their overall perception and use of the tools. The survey largely supported participants’ difficulty in understanding the visualizations and their superficial use of them found in the interview study, but also revealed a path for future work around tutorials and interactive features to support practitioners in using the tools.

“Our next step is to explore ways of helping data scientists form the right mental models so that they can take advantage of the full potential of these tools,” says Wortman Vaughan.

The researchers conclude that as the interpretability landscape continues to evolve, studies of the extent to which interpretability tools are achieving their intended goals and practitioners’ use and perception of them will continue to be important in improving the tools themselves and supporting practitioners in productively using them.

Putting people first

Fairness and interpretability aren’t static, objective concepts. Because their definitions hinge on people and their unique circumstances, fairness and interpretability will always be changing. For Wallach and Wortman Vaughan, being responsible creators of AI begins and ends with people, with the who: Who is building the AI systems? Who do these systems take power from and give power to? Who is using these systems and why? In their fairness checklist and interpretability tools papers, they and their co-authors look specifically at those developing AI systems, determining that practitioners need to be involved in the development of the tools and resources designed to help them in their work.

By putting people first, Wallach and Wortman Vaughan contribute to a support network that includes resources and also reinforcements for using those resources, whether that be in the form of a community of likeminded individuals like in WiML, a comprehensive checklist for sparking dialogue that will hopefully result in more trustworthy systems, or feedback from teams on the ground to help ensure tools deliver on their promise of helping to make responsible AI achievable.

Go to Original Article
Author: Microsoft News Center

Deliver better experiences, insights, and care with Microsoft Cloud for Healthcare – Microsoft Industry Blogs

The world after this pandemic will not be the same as the one that came before it.

From remote teamwork and telehealth, to supply management and customer service, to critical cloud infrastructure and security—we are working alongside customers every day to help manage through a world of remote everything.

The COVID-19 pandemic has impacted nearly every aspect of people’s lives and every aspect of the healthcare system. It’s preventing healthcare delivery practices from operating at normal business levels, it’s disrupting patient access to high-quality medical care, and it’s forcing everyone to think about how to continue pushing forward in new and different ways. Our commitment has always been to ensure the tools we provide are up to the task of supporting our customers in their time of need. Hear CEO Satya Nadella’s words for more on Microsoft’s thoughts for our healthcare workers.

In that same spirit, Microsoft is announcing its first industry-specific cloud offering, Microsoft Cloud for Healthcare, now available in public preview and through a free trial for the next six months. The offer brings together existing and future capabilities that deliver automation and efficiency on high-value workflows, as well as deep data analytics for both structured and unstructured data, that enable customers to turn insight into action. A robust partner ecosystem extends the value of the platform with additional solutions to address the most pressing challenges the healthcare industry is facing today. Healthcare will be the first industry served with additional industry-specific clouds to follow.

Microsoft Cloud for Healthcare brings together trusted and integrated capabilities for customers and partners that enrich patient engagement and connects health teams to help improve collaboration, decision-making, and operational efficiencies. Microsoft Cloud for Healthcare will support accelerated health transformation into the future, with capabilities for customers spanning the most important needs for healthcare organizations:

  • Enhancing patient engagement
  • Empowering health team collaboration
  • Improving operational and clinical data insights
  • Cloud built on interoperability, security, and trust
  • Extensible healthcare partner ecosystem

Enhance patient engagement

More than ever, being connected is critical to creating an individualized patient experience. Microsoft Cloud for Healthcare helps healthcare organizations to engage in more proactive ways with their patients, allows caregivers to improve the efficiency of their workflows and streamline interactions with patients with more actionable results. Organizations can use Microsoft Cloud for Healthcare to extend the value of Microsoft Dynamics 365 Marketing, Dynamics 365 Customer Service, and Azure IoT to deploy:

  • Consumer-friendly patient experience: healthcare organizations can create individualized care plans for patients, or groups of patients, that allows providers to publicize relevant content and proactive outreach to patients on any device when they need it. Deploy secure virtual visits, chatbot assessments, and remote health monitoring to create a connected health experience. One of those tools is our Microsoft Healthcare Bot Service. Since March, more than 1,600 instances of COVID-19 bots based on our service have gone live impacting more than 31 million people across 23 countries. The CDC and healthcare systems from Seattle to Copenhagen, and from Rome to Tel Aviv, are using this service to create COVID-19 self-assessment tools to reduce some of the strain on their emergency hotlines.
  • Connected physician and referral management: care teams can easily create referrals, search for providers, and understand physician spend, satisfaction, and enhanced analytics on referral categories.
  • Enhanced patient engagement portals: patients and providers can easily interact through this self-service portal which enables various healthcare tasks such as online appointment booking, reminders, bill pay, and much more. This also allows providers the ability to engage with patients easily through the device of their choice.
  • Intelligent patient outreach: enables healthcare organizations to design interactive patient journeys to nurture leads, publicize relevant events, and contact patients with preventative and care management programs that help promote better health outcomes.
  • Continuous patient monitoring through IoT: generate secure, scalable data ingestion from medical devices to allow care teams to monitor patients in and outside of clinical facilities. With real-time insights, care teams can provide timely escalations of care, reduce readmissions, and provide personalized, predictive care.

Empower health team collaboration

Even before the current global pandemic, the healthcare industry has been in the midst of a massive shift marked by the rise of team-based care due to increased medical specialization, exponential growth in the volume of digital patient data, and increasingly demanding data protection requirements. Too often, the tools providers use to coordinate patient care are fragmented and impede the collaborative workflows required in a complex care environment.

To address these challenges, we have been building capabilities in Microsoft 365 and Microsoft Teams that streamline healthcare workflows and provide a secure platform for connected care coordination. Teams, which supports HIPAA compliance and is HITRUST certified, brings together chat, voice and video meetings, and offers recording and transcription, as well as secure messaging features, available across devices.

As we’ve seen during the COVID-19 pandemic, clinicians also need greater flexibility and convenience in how they are able to connect with patients. Today we are announcing general availability of the Bookings app in Teams, which enables healthcare providers to schedule, manage, and conduct provider-to-patient virtual visits within Teams.

In an effort to protect patients and providers while maintaining continuity of ambulatory care, healthcare providers including St. Luke’s University Health Network, Stony Brook Medicine, and Calderdale  & Huddersfield NHS Foundation Trust in the U.K. have been using teams to conduct virtual patient visits and provide continuity of care while protecting providers and patients. Patients receive a customized email and can join their appointment in one click on a desktop, or in the Microsoft Teams iOS or Android Mobile apps.

Microsoft Cloud for Healthcare brings together existing and future capabilities important to how care teams communicate, collaborate, coordinate care, and generate insights that help improve patient outcomes and workflow effectiveness.

Improve clinical and operational data insights

Healthcare organizations are taking advantage of building virtual agents, automating workflows, analyzing data, and sharing insights in real-time. COVID-19 has accelerated the urgent need for healthcare organizations to create no-code/low-code apps and workflows in hours or days, not weeks or months. Thousands of organizations are relying on new integration between Microsoft Teams and Power Apps to share timely information.

In just two weeks, Swedish Health Services, the largest non-profit health provider in the Seattle area, used Power Apps to build a solution to track critical supplies.

Microsoft’s newest releases that support FHIR technology enabled the Chicago Dept of Public Health (CDPH) and Rush Hospital with an end to end solution to bring together clinical, lab and capacity data analysis in just a few days to support the COVID crisis. Using the C-CDA converter to FHIR converter, the API for FHIR—a cloud-based FHIR Server, and the Power BI FHIR connector, CDPH was able to ingest data in different formats with a simple API call, convert it to FHIR and add visualizations and downstream analytics that enabled rapid connectivity of data and interoperability across multiple hospitals.

Cloud built on interoperability, security, and trust

Underpinning all these great capabilities is our focus and commitment to interoperability, security, and compliance. We know that to provide the best care, healthcare organizations need to be able to consume, access, and share information rapidly and securely.

The future of highly secure data agility in the cloud– and the interoperability tools that healthcare organizations need to organize their health data in the cloud around FHIR – are integrated into Microsoft Cloud for Healthcare. Last year Microsoft was the first cloud to offer a generally available Azure FHIR service—which allows healthcare organizations to ingest and persist data in the FHIR format.

Healthcare organizations that are already underway with open standards like FHIR have been able to collaborate in rapid time, and it’s given their teams the ability to care for patients with a high level of agility. We’re seeing examples all around the globe and it’s been both humbling and inspiring to see the FHIR technology we’ve built specifically for health workloads in Microsoft Cloud for Healthcare support these efforts.

Security and compliance remain a strategic priority for healthcare organizations, and the shift to remote work only increases the need for integrated, end-to-end security architecture that reduces both cost and complexity. Microsoft has the highest levels of commitment to trust, security, and meeting industry compliance standards and certifications in the industry.

Extensible healthcare partner ecosystem

Microsoft Cloud for Healthcare also enables healthcare systems to take advantage of our robust ecosystem of healthcare partners who can provide solutions that complement and extend core cloud capabilities. Using these partners’ expertise will help organizations through EHR and platform integrations, implementation services, and healthcare SaaS offerings. Microsoft is proud to work closely with the leading providers of health systems, from organizations like Accenture, Adaptive Biotechnologies, Allscripts, DXC Technology, Innovaccer, KPMG, and Nuance to co-develop new solutions with leaders in their respective sectors like Humana, Providence, Novartis, and Walgreens Boots Alliance.

What’s powerful about these tools is that they’re being used not just by providers in the delivery of healthcare but by ISVs, pharmaceutical and life sciences companies, and government systems. We’ve seen companies like KenSci—a healthcare AI & data management platform—launch their Mobile Command Center with Real Time Bed Management, Ventilator Utilization, and Capacity Planning for COVID-19. In just 48 hours they can create a hospital solution—even those using legacy data systems—and help manage their data in the Microsoft Cloud for Healthcare using the open standard of FHIR.

Commitment to industry

We know that technology has a role to play in accelerating progress for solutions to the pandemic and other pressing healthcare concerns and challenges. Looking ahead, we expect to see healthcare organizations continue to use newly implemented technology tools throughout the recovery period and into the new normal. Companies that need to accelerate their digital transformation during this time will continue to rely on trusted partners who can support their trust, security, and technology adoption into the future.

Microsoft Cloud for Healthcare will make it easier for them to remain agile and focus on what they do best – delivering better experiences, insights, and care.

Microsoft Cloud for Healthcare is available for a free trial over the next 6 months. Click to see the launch during the Microsoft Healthcare Summit, view a demo, or learn more about public preview.

Go to Original Article
Author: Microsoft News Center