Designing Machine Learning Systems

Index

Symbols

• 1NF (first normal form), Relational Model
• 2NF (second normal form), Relational Model

A

• A/B testing, A/B Testing-A/B Testing, Bandits
• accuracy-related metrics, Monitoring accuracy-related metrics
• ACID (atomicity, consistency, isolation, durability), Transactional and Analytical Processing
• active learning, Active learning-Active learning
• ad hoc analytics, Experiment tracking
• adaptability, Adaptability
• adversarial attacks, Evaluation challenge
• adversarial augmentation, Perturbation
• AI (artificial intelligence), ethics, Responsible AI, Discover sources for model biases-Discover sources for model biases
  • data-driven approach limitations, Understand the limitations of the data-driven approach
  • irresponsible, case studies, Irresponsible AI: Case Studies-Case study II: The danger of “anonymized” data
  • mitigating biases, Establish processes for mitigating biases
  • model cards, Create model cards-Create model cards
  • trade-offs, Understand the trade-offs between different desiderata
• Airflow, Data Science Workflow Management-Data Science Workflow Management
• alert fatigue, Monitoring features, Alerts
• alert policies, Alerts
• algorithms
  • bandit algorithms, Bandits-Contextual bandits as an exploration strategy
  • continual learning and, Algorithm challenge-Algorithm challenge
  • feature importance, Feature Importance
• analytical processing, Transactional and Analytical Processing
• Apache Iceberg, Transactional and Analytical Processing
• architectural search, Hard AutoML: Architecture search and learned optimizer
• Argo, Data Science ...