Page 11 - MSDN Magazine, December 15, 2017

Trained Services
Cognitive Services
AI On Data
Custom Services
Azure Machine Learning
Azure ML Studio
Azure ML Workbench
Azure Notebooks
VS Code Tools for AI
Data Lake
SQL Server
Cosmos DB
Spark
DVSM Batch AI ACS
Cognitive Toolkit
TensorFlow
Caffe 2
Figure 1 The Microsoft AI Platform
AI SERVICES Conversational AI
BOT Framework
AI TOOLS
Microsoft AI Platform
The Microsoft AI platform provides a suite of powerful tools, such as the Bot Framework, Cognitive Services, Azure Machine Learning and many more. These tools allow developers to easily and quickly infuse AI into their applications and scenarios, enabling new, intelligent experiences for their users.
Powered by the enterprise-ready capabilities of Azure, the Microsoft AI platform presents a rich set of
AI INFRASTRUCTURE
AI On Data
Deep Learning Frameworks
• Retail: Lowe’s partnered with Microsoft to create a kitchen remodeling design experience driven by AI. The Lowe’s cus- tomer shares his or her dream kitchen photos with a design specialist, who uses an AI-powered application to gain deep insight into the style and preferences of the customer. These findings are used to generate a match from the Lowe’s dream kitchen collection, which is then shown in a mixed-reality environment so the customer can interact with the products (such as cabinets, counter tops and appliances). Like Lowe’s, many retailers are turning to AI to re-imagine the retail experience for consumers.
interoperable services, APIs, libraries, frameworks and tools that developers can leverage to build smart applications. The Microsoft AI platform consists of three core areas: AI Services, AI Infrastruc- ture and AI Tools. Let’s start with a quick tour of the AI platform, shown in Figure 1.
Cognitive Services: Trained services like Cognitive Services enable you to jumpstart development of your AI applications, with- out requiring you or your data science team to develop and train the models. Cognitive Services features a rich set of instant AI capabil- ities that you can use. These AI capabilities are organized into the following categories: vision, speech, language, knowledge and search.
Tohelpeliminatetheheavyliftinginvolvedinbuildingend-to-end systems, Microsoft provides a powerful AI platform, composed of a set of loosely coupled but highly interoperable services on Azure.
Building these kinds of AI applications requires the integration of numerous components. To help eliminate the heavy lifting involved in building end-to-end systems, Microsoft provides a pow- erful AI platform, composed of a set of loosely coupled but highly interoperable services on Microsoft Azure. This article provides an overview of this platform and points to valuable resources for getting started developing exciting AI applications.
The Cognitive Services APIs enable you to leverage powerful computer vision algorithms that have been pre-trained to recognize things like different face attributes, landmarks, celebrities, gender, emotion, and printed or written words (Optical Character Rec- ognition, or OCR). Powerful language capabilities can recognize commands from users, analyze key phrases, perform translations and spell check, and more.
Figure 2 Using CustomVision.AI to Build a Custom Travel Attractions Computer Vision Model msdnmagazine.com
Customized Computer Vision Models: As you explore Cogni- tive Services to develop your AI applications, you may find that you need to further customize the models using your own data. You can do that with Custom Vision services (customvision.ai). Custom Vision lets you bring your own data, and use it to train your com- puter vision models. Underneath the hood, state-of-the-art transfer learning techniques leverage existing pre-trained computer vision models, and evolve them to learn about the new images you
Dec. 15, 2017 / Connect(); Special Issue 7