Test AI-900 Scope of Knowledge

Overall, AI-900 exam tests the applicant's understanding of AI and ML through various topics categorized into 5 domains:

• Workloads for Natural Language Processing (NLP) related to Azure (15-20%)

  The fourth portion of AI-900 exam deals with Natural Language Processing or NLP workloads in Azure. To get a passing score, the entrant must have a firm grasp on the facets and usage of diverse NLP scenarios for workloads which comprise the extraction of the key phrase, the recognition of entities, the analysis of sentiments, modeling for languages, the recognition of speech, and translation traits. The applicants are also required to know about different techniques as well as services of Azure used for various NLP workloads. They are required to be capable of identifying diverse abilities of various services such as text analysis, LUIS, speech, and translator text.

• Fundamental principles of machine learning on Azure (30-35%)

  In the second section, the applicant's knowledge of fundamental principles of machine learning in relation to Azure cloud computing services will be tested. This requires the applicant to have a clear comprehension of all the common types of various scenarios for ML that include classification, regression, and clustering. It is also essential to have a thorough understanding of the core MI concepts such as labels & features in datasets, dataset training that go with the validation process, etc. The entrant must as well as be capable of identifying the core task required in developing an ML solution. So, understanding the ingestion and preparation of data chunks, selection and engineering of various features, evaluation procedure as well as training for models, and deployment & management of those models will be vital. Finally, this portion of AI-900 exam includes concepts of ML automated user interface that are ML Wizard UI and Azure ML designer.

• AI workloads and considerations (15-20%)

  The first topic of the final exam deals with AI workloads that are usually common. It includes candidates' understating various types of workloads such as predicting forecasting traits, the detection of anomalies, vision for computer systems, NLP, and conversational AI workloads. The examinees are also required to have a thorough understanding of the principles of AI that possess responsibility traits. It means that the candidate needs to be well aware of all the considerations for different solutions for AI which include fairness, reliability, security, privacy, inclusiveness, transparency facets, and finally, functions for accountability.

• Conversational workloads for AI based on Azure and features (15-20%)

  The fifth and final domain for such an exam tests the entrant's understanding of conversational AI workloads. This requires the comprehension of facets as well as cases for the usage of conversational traits for AI which includes web chatbots, voice menus for the telephone, and digital assistants used for personal purposes. The applicant also needs to know how to check various Azure services related to conversational AI. Additionally, they should have a firm understanding of the QnA Maker Service & Bot service. Finally, it is essential for candidates to know and be able to describe the common characteristics of conversational artificial intelligence solutions.

• Workloads for computer vision based on Azure and features (15-20%)

  This segment of the official test is all about workloads for computer vision in relation to Azure services. To get through this part, one must have a general grasp of the various common forms of computer vision solutions that include image classification, object detection, semantic segmentation, recognition for the optical character, and facial detection. The applicants are also required to have an understanding of the diverse methods including services used in Azure for activities that utilize computer vision. Such methods and services include the vision for computers, vision for custom, services for face vision, and at last, the recognizer of the form.

Reference: https://docs.microsoft.com/en-us/learn/certifications/exams/ai-900

About Exam AI-900

The interested candidate for the Microsoft AI-900 test, which also goes under the name the Azure test for AI Fundamentals, should have at least some foundational understanding of the concept of AI features alongside ML ones. This exam will test the entrant's knowledge of different AI common workloads, principles of AI and ML, NLP workloads, and other related concepts. When it comes to the details for this exam, the Microsoft AI-900 is available in several languages, among which you'll find Spanish, English, Korean, German, Chinese (Simplified), and French. Such an exam comes with a registration fee of $99 and should be scheduled through Pearson VUE. Also, in the main exam, candidates will be exposed to around 40-60 tasks with 1 hour given to solve them.