AI being among the relatively primitive technologies, that most of the advance AI services out there are so end-to-end connected, that the end users sometimes don’t even realize that there is AI in process. There are so few options out there that provide a platform for the end user to implement AI for a custom purpose. Microsoft offers a few services in the AI field for all ranges of users (entry-level to advance), with Microsoft Azure Machine Learning Studio being the core product, it offers several services like Cognitive Services, Azure Bot Service, Azure Cognitive Search and Data Science Virtual Machines. Let us look at Microsoft Azure Machine Learning Studio, and Data Science Virtual Machines in more detail.
Microsoft Azure Machine Learning Studio
Microsoft Machine Learning Studio is a very intuitive and caters to the needs of people of various skill levels. It is a very beginner friendly tool, especially for people who had previously worked on Data Integration Platforms like Informatica, snapLogic, Jitterbit. With an intuitive GUI, a user can very easily drag-and-drop datasets and analysis modules to experiment with different models to observe different behaviors. This is especially useful, since there is not a universally agreed upon model that will run correctly with a given dataset and use case, there is a need for one to implement and perform several different models before they discover the most optimal for the purpose. A visual representation of the model helps non-technical people get a better overview of a process than a typical Python or R code.
While it caters very well to the amateur developers and non-technical people involved a project, it also allows several Python and R Modules that allow advance developers to create scripts in their familiar language to perform said task, including visualization. Making it the perfect tool for teams and projects with diverse skill.
Data Science Virtual Machines
Data Scientists and Machine Learning Engineers and Enthusiasts know about the significant improvement in processing when implementing deep learning on Graphics Processing Unit (GPU) accelerated machines, and the struggle that comes with setting up a GPU enabled machine for the use and installing compatible libraries. Microsoft provides Data Science Virtual Machines that come with pre-installed GPU drivers, and GPU versions of the libraries and frameworks that are needed for several ML and DL applications, as an option. Even without the GPU, a DSVM is a very handy service for those who would like to implement ML projects on the cloud.
