Building Machine Learning Powered Applications. Going from Idea to Product

Kup polskie wydanie:

Spis treści ebooka

• Preface
  • The Goal of Using Machine Learning Powered Applications
    • Use ML to Build Practical Applications
    • Additional Resources
  • Practical ML
    • What This Book Covers
      • The entire process of ML
      • A technical, practical case study
      • Real business applications
    • Prerequisites
    • Our Case Study: MLAssisted Writing
    • The ML Process
  • Conventions Used in This Book
  • Using Code Examples
  • OReilly Online Learning
  • How to Contact Us
  • Acknowledgments
• I. Find the Correct ML Approach
• 1. From Product Goal to ML Framing
  • Estimate What Is Possible
    • Models
      • Classification and regression
      • Knowledge extraction from unstructured data
      • Catalog organization
      • Generative models
    • Data
      • Data types
      • Data availability
      • Datasets are iterative
  • Framing the ML Editor
    • Trying to Do It All with ML: An End-to-End Framework
    • The Simplest Approach: Being the Algorithm
    • Middle Ground: Learning from Our Experience
  • Monica Rogati: How to Choose and Prioritize ML Projects
  • Conclusion
• 2. Create a Plan
  • Measuring Success
    • Business Performance
    • Model Performance
    • Freshness and Distribution Shift
    • Speed
  • Estimate Scope and Challenges
    • Leverage Domain Expertise
      • Learning from experts
      • Examining the data
    • Stand on the Shoulders of Giants
      • Open data
      • Open source code
      • Bring both together
  • ML Editor Planning
    • Initial Plan for an Editor
    • Always Start with a Simple Model
  • To Make Regular Progress: Start Simple
    • Start with a Simple Pipeline
      • Training
      • Inference
    • Pipeline for the ML Editor
  • Conclusion
• II. Build a Working Pipeline
• 3. Build Your First End-to-End Pipeline
  • The Simplest Scaffolding
  • Prototype of an ML Editor
    • Parse and Clean Data
    • Tokenizing Text
    • Generating Features
  • Test Your Workflow
    • User Experience
    • Modeling Results
      • Finding the impact bottleneck
  • ML Editor Prototype Evaluation
    • Model
    • User Experience
  • Conclusion
• 4. Acquire an Initial Dataset
  • Iterate on Datasets
    • Do Data Science
  • Explore Your First Dataset
    • Be Efficient, Start Small
    • Insights Versus Products
    • A Data Quality Rubric
      • Data format
      • Data quality
      • Data quantity and distribution
      • ML editor data inspection
  • Label to Find Data Trends
    • Summary Statistics
      • Summary statistics for ML editor
    • Explore and Label Efficiently
      • Vectorizing
        • Tabular data
        • Text data
        • Image data
      • Dimensionality reduction
      • Clustering
    • Be the Algorithm
    • Data Trends
  • Let Data Inform Features and Models
    • Build Features Out of Patterns
      • Raw datetime
      • Extracting day of week and day of month
      • Feature crosses
      • Giving your model the answer
    • ML Editor Features
  • Robert Munro: How Do You Find, Label, and Leverage Data?
  • Conclusion
• III. Iterate on Models
• 5. Train and Evaluate Your Model
  • The Simplest Appropriate Model
    • Simple Models
      • Quick to implement
      • Understandable
      • Deployable
    • From Patterns to Models
      • We want to ignore feature scale
      • Our predicted variable is a linear combination of predictors
      • Our data has a temporal aspect
      • Each data point is a combination of patterns
      • ML Editor model
    • Split Your Dataset
      • Validation set
      • Test set
      • Relative proportions
      • Data leakage
        • Temporal data leakage
        • Sample contamination
    • ML Editor Data Split
    • Judge Performance
      • Bias variance trade-off
      • Going beyond aggregate metrics
  • Evaluate Your Model: Look Beyond Accuracy
    • Contrast Data and Predictions
    • Confusion Matrix
    • ROC Curve
    • Calibration Curve
    • Dimensionality Reduction for Errors
    • The Top-k Method
      • The k best performing examples
      • The k worst performing examples
      • The k most uncertain examples
      • Top-k implementation tips
      • Top-k method for the ML Editor
    • Other Models
  • Evaluate Feature Importance
    • Directly from a Classifier
    • Black-Box Explainers
  • Conclusion
• 6. Debug Your ML Problems
  • Software Best Practices
    • ML-Specific Best Practices
  • Debug Wiring: Visualizing and Testing
    • Start with One Example
      • Visualization steps
      • Data loading
      • Cleaning and feature selection
      • Feature generation
      • Data formatting
      • Model output
      • Systematizing our visual validation
      • Separate your concerns
    • Test Your ML Code
      • Test data ingestion
      • Test data processing
      • Test model outputs
  • Debug Training: Make Your Model Learn
    • Task Difficulty
      • Data quality, quantity, and diversity
      • Data representation
      • Model capacity
    • Optimization Problems
  • Debug Generalization: Make Your Model Useful
    • Data Leakage
    • Overfitting
      • Regularization
      • Data augmentation
      • Dataset redesign
    • Consider the Task at Hand
  • Conclusion
• 7. Using Classifiers for Writing Recommendations
  • Extracting Recommendations from Models
    • What Can We Achieve Without a Model?
      • Using feature statistics
    • Extracting Global Feature Importance
    • Using a Models Score
    • Extracting Local Feature Importance
  • Comparing Models
    • Version 1: The Report Card
    • Version 2: More Powerful, More Unclear
    • Version 3: Understandable Recommendations
  • Generating Editing Recommendations
  • Conclusion
• IV. Deploy and Monitor
• 8. Considerations When Deploying Models
  • Data Concerns
    • Data Ownership
    • Data Bias
      • Test sets
    • Systemic Bias
  • Modeling Concerns
    • Feedback Loops
    • Inclusive Model Performance
    • Considering Context
    • Adversaries
      • Defeating a model
      • Exploiting a model
    • Abuse Concerns and Dual-Use
  • Chris Harland: Shipping Experiments
  • Conclusion
• 9. Choose Your Deployment Option
  • Server-Side Deployment
    • Streaming Application or API
    • Batch Predictions
  • Client-Side Deployment
    • On Device
    • Browser Side
  • Federated Learning: A Hybrid Approach
  • Conclusion
• 10. Build Safeguards for Models
  • Engineer Around Failures
    • Input and Output Checks
      • Check inputs
      • Model outputs
    • Model Failure Fallbacks
      • Filtering model
  • Engineer for Performance
    • Scale to Multiple Users
      • Caching for ML
        • Caching inference results
        • Caching by indexing
    • Model and Data Life Cycle Management
      • Reproducibility
      • Resilience
      • Pipeline flexibility
    • Data Processing and DAGs
  • Ask for Feedback
  • Chris Moody: Empowering Data Scientists to Deploy Models
  • Conclusion
• 11. Monitor and Update Models
  • Monitoring Saves Lives
    • Monitoring to Inform Refresh Rate
    • Monitor to Detect Abuse
  • Choose What to Monitor
    • Performance Metrics
    • Business Metrics
  • CI/CD for ML
    • A/B Testing and Experimentation
      • Choosing groups and duration
      • Estimating the better variant
      • Building the infrastructure
    • Other Approaches
  • Conclusion
• Index

pokaż cały spis treści

ukryj recenzje