Azure AI Fundamentals - Computer Vision and Machine Learning
Cloud computing meets AI
This is a story to talk about how Artificial Intelligence has combined with the cloud computing service offered by Microsoft called Azure. Artificial intelligence allows for improvements in fields like health care and empower smart infrastructure as well as help with entertainment experiences & Azure allows developers to use the full power of AI from a cloud service. Artificial intelligence, in the simples of terms, is the creation and designing of software programs that can imitate human capabilities. There are many key components of AI, including machine learning (a way to 'teach' a computer how to make conclusions from data and predict events/patterns), anomaly detection (the ability of a computer program to automatically detect errors or unusual activity in a system), computer vision (capability of software to interpret the world visually using cameras, video, and pictures), natural language processing (the power of a software program to understand and interpret written or spoken language and respond), and conversational AI (ability of software to participate in a conversation). Machine learning is the foundation for most, if not all, AI solutions being implemented with Azure. It allows AI to be used in a way that can help deal with many real-world problems. Azure and Microsoft have created videos about how their AI systems was used to create a program in Australia called The Yield that helps farmers make informed decisions related to weather, soil and plant conditions. This is only one of the many types of programs that people can create with AI and Azure for the benefit of society. Machine learning collects data to train machines to make decisions using pre-existing information. The data can be static or collected continuously using sensors from homes, cars, cities, public transport infrastructure, and factories.
Within Azure, you have a main service known as Azure Machine Learning service - a cloud-based system for creating, using, and publishing machine learning models. This service has multiple features and capabilities for users of Azure including automated machine learning (a system that allows non-experts to quickly create an effective machine learning model from data), Azure Machine Learning designer ( a basic graphical interface that allows no-code creation of machine learning solutions), data and compute management (cloud-based data storage that data scientists can use to run experiment code at scale), and pipelines (ways to orchestrate model training, deployment, and management tasks). These four resources can make it easy for experts and the unprepared to create proper machine learning solutions.
Azure also provides anomaly detection - a machine learning technique that looks at data over times to identify any kind of unusual changes in the information or computer systems. For this, Azure has the Anomaly Detector service which provides an application programming interface (API) that developers can use to create anomaly detection solutions. It can used if you are doing something like creating a form of financial software to monitor credit card transactions to find fraud or a system in a race car to warn the driver/engineers about potential mechanical failures before they can completely cripple the vehicle.
Computer Vision is part of Azure and was used by a team that developed the Seeing AI application to help the blind and those with low vision. The app describes to them nearby people, text, and objects; this program basically acts as the narrator to help those who are vision-impaired navigate the world around them. Computer vision solutions can be used to classify images, detect objects, identify objects, analyze images, and recognize individuals. Azure provides Computer Vision, Custom Vision (a customizable version of computer vision), Face (a service for facial recognition), and form recognizer (extracts forms and invoices).