The ultimate guide to Microsoft Azure AI Fundamentals (AI 900)
Anyone interested in AI and ML must have come across Azure AI fundamentals. This blog is all about it. Let’s dive in.
About Microsoft Azure AI 900
Azure AI-900 is a fundamental exam conducted by Microsoft Azure on AI and ML. This exam tests the candidate’s knowledge of Artificial Intelligence (AI), Machine Learning (ML), and related Microsoft Azure services. It’s the first step in achieving role-based certifications in Azure, such as Azure AI Engineer Associate or Azure Data Scientist Associate. It’s an excellent opportunity to enhance one’s career prospects and demonstrate knowledge of AI and ML workloads and related Azure services.
Prerequisites for the exam
- A good understanding of the fundamentals of Machine Learning and Artificial Intelligence concepts.
- Basic knowledge of Azure
Syllabus
Basic knowledge in the following topics:
- Machine Learning Algorithms and Classification
- AI Fundamentals
- Computer Vision
- Natural Language Processing
- Document Intelligence and Knowledge mining
- Generative AI
Type of Questions Asked
- Multiple Choice
- True or False
- Drag and Drop
- Hot Spot
Prep Time
The required preparation time varies based on individual knowledge. An individual completely new to AI may take a week to prepare. However, for those familiar with basic concepts and how things work in ML, a day or two is enough.
Study Materials
Microsoft provides courses to help candidates prepare. There are two types of courses: Self-paced and Instructor-led.
- Self-paced training allows candidates to learn on their own without any timeframe in their personalized learning paths.
- Instructor-led training provides dedicated personal attention from technical subject matter experts.
I chose self-paced learning, and I believe it suffices. Learning the study materials provided by Microsoft and going through some question papers is all one would need to pass the exam with flying colors.
Important Concepts
- Principles of Responsible AI
- Fairness
- Reliability and Safety
- Privacy and Security
- Inclusiveness
- Transparency
- Accountability
2. Applications of Machine Learning Algorithms
- Supervised — Prediction/labeled data to known output
- Unsupervised — Classification/identifying unknown patterns
- Reinforcement — Trial and Error method [Reward/Punishment based]
3. Confusion Matrix
- Metrics used for Classification (True Positive Rate / ROC Curve) and Regression (Mean Absolute Error, Root Mean Squared Error, Relative Squared Error, Relative Absolute Error, Coefficient of Determination)
- Model Evaluation — The Model evaluation module outputs a confusion matrix showing the number of true positives, false negatives, false positives, and true negatives, as well as ROC, Precision/Recall, and Lift curves.
- Feature Engineering — Using existing features to create new features that help ML algorithms learn better. New features are derived from existing ones.
- Feature Selection — Selecting a subset of relevant, useful features to use in building an analytical model.
4. Computer Vision
- Difference between Image Classification, Object Detection, Semantic Segmentation, Image Analysis, Face Detection, Face Analysis, Face Recognition and Optical Character Recognition
5. Natural Language Processing
- Understanding of Entity recognition, speech recognition, speech synthesis, Language Understanding Intelligent Service (LUIS) and Key phrase extraction.
6. Understanding of QnA Maker and Azure Bot Service.
Other Resources that might be helpful
I believe with these resources one can easily pass the exam with flying colours. All the very best to everyone writing the exam.
Good Luck!
