Publisher's Synopsis
Valliappa Lakshmanan's Learning About Machine And Designing All Patterns - The book's design patterns capture best practices and solutions to recurring machine learning problems. The authors, three engineers from Google, list best practices to help data researchers solve common problems that arise during the ML process. These design templates incorporate the experiences of hundreds of experts into simple, accessible advice. In this book you will find a detailed explanation of 30 models of data representation and problems, operationalization, repeatability, repeatability, flexibility, explainability and fairness. Each sample includes a description of the problem, a number of possible solutions, and recommendations for selecting the appropriate technique for your situation.In This Book You Will Learn: -Identify and mitigate common challenges in training, evaluating and implementing ML models-Representation of data for different types of ML models, including embedding, feature crossovers, etc.-Select the type of model appropriate for your specific problems-Create a robust training cycle that uses checkpoints, distribution strategy, and tuning hyperparameters-Install scalable ML systems that you can recycle and update to reflect new data-Interpret model predictions for stakeholders and ensure models treat users fairly
