A montage of recent customer use cases, demonstrating the range of capabilities of Microsoft’s AI, Big Data & Advanced Analytics offerings.
Adobe Helps Customers Break the Language Barrier
Adobe Experience Manager (AEM) is an enterprise content management platform that helps large and mid-size companies manage their web content, digital assets, online communities and more. Digital marketers use AEM to manage and personalize online content, including user-generated content on community sites. AEM helps companies build their brands online, driving demand and expanding their markets.
Companies using AEM often serve customers in many countries and languages, and customers were looking to Adobe for an easy way for their digital marketers to translate multinational web sites and content from international user communities into other languages. The sheer volume of information being posted online and prohibitive costs of human translation services meant that most companies could not afford the time and budget needed for manual translation efforts.
Adobe decided to address the problem by integrating AEM with Microsoft Translator. The Microsoft research team behind the Translator technology had spent over a decade to build a linguistically informed, statistical machine translation service that learns from prior translation efforts. What’s more, the service was flexible, reliable and performant at a massive scale, having performed billions of translations daily, and all while respecting customer data privacy.
To make the system work, all that Adobe’s customers had to do was supply the system with a set of parallel documents, containing the same information in two languages (i.e. the source and target languages). The system would then analyze this material and build a statistical model layered on top of Translator’s generic language knowledge. Translator’s statistical models and algorithms allow the system to automatically detect correlations between source and target language in the training data, helping it determine the best translation of a new input sentence.
AEM versions 6.0 and above now ship with a pre-configured Translator connector, and include a free trial license of 2 million characters per month, enabling users to start automatic translation with minimal effort. Adobe customers can now build, train, and deploy customized translation systems that understand a preferred terminology or style that is specific to their industry or domain. Some customers are also using this new capability to further expand their audience, providing content in previously underserved languages where there is a latent demand or audience. You can learn more about the Adobe solution here.
As people all over the world increasingly look to online communities for instant answers and information, they do not want to be restricted by their language skills. By partnering with Microsoft, Adobe is helping customers become productive by breaking through language barriers.
Dallas Zoo Tracks Elephant Behavior with Fitness Bands
Zoos worldwide have been working hard to provide larger, more natural and varied environments for their elephants. The Dallas Zoo, however, decided to go an extra step further: A growing number of elephants at the Dallas Zoo are now part of a pioneering application that uses RFID bracelets and Microsoft Azure to better understand elephant behavior and provide more customized care.
The earlier solution deployed by the Zoo relied on a combination of video cameras and direct observation by staff to track the animals, but that approach left big gaps, and even caused occasional errors in the data. What’s more, the Zoo had to manage unwieldy spreadsheets in the past, to understand their animals’ movements. Since their software could handle only 15 days’ worth of data at a time, that made insights from long-term data (such as behavioral changes in an elephant as it aged) near impossible. The earlier system also made it impossible to integrate observations with external data such as weather changes, fluctuations in the zoo’s attendance, and the like.
The introduction of elephant “ankle bracelets” powered by RFID technology has decidedly changed things for the better. Since elephants are trained to show their feet to handlers for exams and pedicures, getting the bands on them is simple, quick, and stress-free. The Zoo is now able to track a very wide range of parameters for each animal. Nancy Scott, the Coordinator of Elephant Behavioral Science at the Dallas Zoo, now knows that her elephants each walk an impressive average of 10 miles a day. She also knows that Congo, whom she’s dubbed “The Great Explorer”, can walk nearly 17 miles a day, and is also the first one out of the gate when given access to an adjacent habitat to mingle with other species such as giraffe, zebra, and ostrich. That’s useful information to measure the health of elephants not only against their own histories but also against the typical range of the herd. Scott now knows where in the five-acre Giants of the Savanna exhibit the elephants like to go, and where they don’t. She knows who’s been frequenting the mud wallows, pools, scratching posts, log piles or shady spots. She also has a keener understanding of whether the elephants have enough space and how they are using that space, so she can help devise ways to optimize their use of the exhibit.
Elephants being highly social, this technology is also helping the zoo better understand their interactions. They can see which elephants are loners (keeping their distance) vs. which ones are potential friends (frequently traveling together or stationary at night together). When an elephant suddenly moves more slowly or stays near one spot, Scott knows the animal may be ill, leading to faster diagnoses and better health outcomes.
US Medical IT, a Microsoft solution provider and part of the startup community at the University of Texas at Dallas (UTD) Venture Development Department, was an important partner in this effort. With the UTD Venture Development Center’s financial support and US Medical IT’s expertise, the Dallas Zoo enhanced its RFID system with key components of the Microsoft cloud. A SQL Server 2016 -based data warehouse hosted on Microsoft Azure synchronizes the RFID data daily and links it to five other data sources. The data is then made available to Power BI for analysis and to other reporting services running in Azure. The results of the analyses are displayed on dashboards on PCs and mobile devices, including on Apple watches, making insights available to handlers working in the exhibits, to visitors using proposed information kiosks, and to Scott, no matter where she happens to be.
The Zoo can now collect and analyze data across multiple years rather than just days. Information for additional internal and external data sources such as weather, zoo attendance, moon cycles and more can be factored into the analyses. And the best part is that staffers avoid the need for the setup and maintenance of computer systems. The success of the solution has already led Scott to consider ways to expand it further. The addition of Azure Machine Learning, for example, can enable the Zoo to anticipate their elephants’ future needs. The technology can be expanded to other animals, including giraffes, ostriches, and zebras. Gorillas and other apes and monkeys pose an interesting challenge to the technology, because they also move in a third dimension, when they climb trees or other structures. Scott is interested in exploring how their solution could be enhanced to take that into account. You can learn more about the Dallas Zoo solution here.
Institutions around the world involved in the area of animal care are getting inspired by the Dallas Zoo’s pioneering work.
Financial Fabric Helps Hedge Funds Leverage Big Data Analytics Securely, in the Cloud
Since the financial crisis of 2008, the cost of managing hedge funds has grown in pace with the increased regulatory requirements that funds are now expected to meet. According to Preqin, a data provider for the alternative assets industry, manually gathering data from disparate silos, analyzing information, and creating reports can eat up more than 70 percent of a small or midsize hedge fund’s operating budget. On top of that, any discrepancies arising from manual data-handling processes and disjointed workflows can leave a fund vulnerable to regulatory penalties and ultimately to the loss of business.
Hedge funds used to handle many of their IT responsibilities in-house in the past, but it is increasingly clear to modern fund operators that they need much more robust IT infrastructure, including data platforms with advanced analytics capabilities and strong cybersecurity protection, and best practices for security, regulatory compliance and portfolio management.
Financial Fabric is a company that offers hedge funds, institutional investors, and other financial organizations a centralized way to store, analyze and report investment data, using cloud services. To meet the needs of their customers, Financial Fabric required a technology platform that enables fund managers to make data-driven investment decisions without compromising security and privacy. The company decided to base their DataHub solution on Microsoft Azure, taking advantage of its many security features, including the ‘Always Encrypted’ capability of Azure SQL Database.
DataHub includes a client-dedicated data warehouse that ingests information from multiple service providers and systems including prime brokers, fund administrators, order management systems, and industry data sources. In the past, analysts typically downloaded files and documents manually in various formats, and then painstakingly gathered the information into spreadsheets and other tools. Instead, in the new solution, information from diverse sources is automatically collected, cleansed, normalized and loaded to the DataHub, providing up-to-date and accurate information in one place.
Encrypted and stored in the cloud, the information is continuously available to a hedge fund’s business users – including portfolio managers, risk managers, analysts, chief operations officers, and chief financial officers – through business intelligence tools connected to Microsoft SQL Server Analysis Services. Working with interactive Microsoft Power BI dashboards in Excel workbooks, Financial Fabric’s data science team can securely collaborate with clients and create analytics and reports. The DataHub also enables clients to quickly and easily create custom analytics and reports themselves, without IT help. They can also automate workflows such as reconciling data across trades, investment holdings or positions, and cash and margins.
Accessible from virtually anywhere, the analytics and reports are hosted on a Microsoft SharePoint Server farm running on Azure VMs. Historically, the ability to share information on demand while keeping it secure has been an elusive goal for investment managers, but, with DataHub, they are able to securely share information with clients and data scientists to build a more data-driven business with significantly lowered risk. Financial Fabric uses Azure Active Directory and Azure Multi-Factor Authentication to control access throughout all layers of DataHub. The built-in security capabilities have played a critical role in boosting the financial sector’s confidence in cloud solutions.
Financial Fabric’s primary goal was to solve a business challenge, not deal with technical issues. With Azure, the company and its clients can get on the fast track to data science and avoid spending months and millions of dollars buying or creating software. “We have more data scientists on our team than software developers,” says Subhra Bose, CEO at Financial Fabric. “And they’re focused on the clients’ data, calculating things like investment performance and risk exposure. We have also completely separated our platform development from the data analytics on Azure. That’s given us a tremendous amount of mileage, because we can onboard a client without writing a single line of code.”
One of Financial Fabric’s customers, Rotation Capital, chose to bypass the traditional application-specific approach to developing an institutional infrastructure in favor of DataHub. Within a month, the firm gained a powerful, highly secure data platform with minimal investment in IT staff, software, servers, and other operational overhead. Biagio Iellimo, Controller at Rotation Capital, notes, “Software implementation in the hedge fund industry is a huge pain point. Implementations traditionally take anywhere from six months to a year and a half. So the fact that we were up and running on the Financial Fabric DataHub platform within four weeks is beyond impressive.”
DataHub is a cost-effective, scalable and flexible solution that’s helping hedge funds like Rotation Capital take advantage of big data analytics and protect confidential information while meeting ever-changing business requirements. You can read more about the Financial Fabric solution here.
The availability of secure, cloud-based analytics is proving to be fundamentally transformative for the financial services industry.
CIML Blog Team
from Cortana Intelligence and Machine Learning Blog https://blogs.technet.microsoft.com/machinelearning/2017/01/11/how-adobe-dallas-zoo-and-financial-fabric-are-using-microsoft-cognitive-services-and-the-azure-data-platform/