Please enable JavaScript.

Coggle requires JavaScript to display documents.

AI-900 - Coggle Diagram

- - - - Identity potential harms
      - Measure the presence of these harms in the generated outputs
      - Mitigate the harms at multiple layers in your solution to minimize their presence
      - Operate the solution responsibly
- - - - Named entity recognition
      - Sentiment analysis
      - Returned values
        
        ISO 6391 Code
        
        Score
        
        Language Name
      - Universal Language Model (ULM)
        
        speech-to-text
        
        Scenarios
        
        Dictation
        
        Conversational
      - Features
        
        Sentiment analysis
        
        Entity Linking
        
        Personally Identifiable Information (PII) detection
      - QnA Maker
      - Language detection
      - Key phrase extraction
    - - data sources
        
        web page
        
        existing FAQ document
        
        manually entered data
  - - - Specialized models
        
        landmarks
        
        celebrities
        
        brand detection
      - Features
        
        object detection
        
        bounding box coordinates
        
        Class name (ex. apple, orange)
        
        Probability score
        
        facial recognition
        
        tagging
        
        content organization
        
        categorizing
        
        OCR
        
        spatial analysis
      - Custom Vision Model performance metrics
        
        Average Precision
        
        Recall
        
        Precision
      - Form Recognizer
        
        Custom model
        
        Pre-build receipt model
    - - facial analysis / face attributes
        
        exposure
        
        noise
        
        occlusion
      - face matching
        
        face identification
        
        face verification
      - facial detection
      - facial recognition
- - - - Metrics
        
        Recall
        
        Accuracy
        
        Precision
        
        F1 Score
        
        Selectivity
        
        AUC (Area Under Curve)
    - - Labels (function results) and Features (function parameters)
      - Multiple linear regression model
      - Evaluation metrics
        
        Mean absolute error (MAE)
        
        Root mean squared error (RMSE)
        
        Relative absolute error (RAE)
        
        Relative squared error (RSE)
        
        Coefficient of determination (R2)
  - - - Dataset
      - Workspace
    - - Azure Kubernetes Service cluster
      - Azure container instance
    - - Job
      - Compute cluster
    - - Classification
      - Regression
      - Time-series forecasting
    - - Feature selection
      - Finding and removing data outliers
      - Normalize numeric values
      - Impute missing values
    - - Designer
      - Automated ML
      - Notebooks
  - - - Import dataset
      - Select columns in dataset
      - Clean missing data
      - Split data
      - Choose the model (untrained/pretrained)
      - Train the model
      - Score model
      - Evaluate model