Machine Learning. Design Patterns

The design patterns presented in the book reflect the best practical approaches to solving typical machine learning problems. The indicated patterns, implemented in code, concentrate the experience of hundreds of experts into simple and accessible advice. The book contains a detailed analysis of 30 patterns that serve for data and task representation, model training, fault-tolerant service, reproducibility, and artificial intelligence. Each pattern includes a problem statement, a range of potential solutions, and recommendations for selecting the technical approach best suited to the situation.