Böcker av Pete Warden

Deep learning networks are getting smaller. Much smaller. The Google Assistant team can detect words with a model just 14 kilobytes in size—small enough to run on a microcontroller. With this practical book you'll enter the field of TinyML, where deep learning and embedded systems combine to make astounding things possible with tiny devices.Pete Warden and Daniel Situnayake explain how you can train models small enough to fit into any environment. Ideal for software and hardware developers who want to build embedded systems using machine learning, this guide walks you through creating a series of TinyML projects, step-by-step. No machine learning or microcontroller experience is necessary.Build a speech recognizer, a camera that detects people, and a magic wand that responds to gesturesWork with Arduino and ultra-low-power microcontrollersLearn the essentials of ML and how to train your own modelsTrain models to understand audio, image, and accelerometer dataExplore TensorFlow Lite for Microcontrollers, Google's toolkit for TinyMLDebug applications and provide safeguards for privacy and securityOptimize latency, energy usage, and model and binary size

To help you navigate the large number of new data tools available, this guide describes 60 of the most recent innovations, from NoSQL databases and MapReduce approaches to machine learning and visualization tools. Descriptions are based on first-hand experience with these tools in a production environment.This handy glossary also includes a chapter of key terms that help define many of these tool categories:NoSQL DatabasesDocument-oriented databases using a key/value interface rather than SQLMapReduceTools that support distributed computing on large datasetsStorageTechnologies for storing data in a distributed wayServersWays to rent computing power on remote machinesProcessingTools for extracting valuable information from large datasetsNatural Language ProcessingMethods for extracting information from human-created textMachine LearningTools that automatically perform data analyses, based on results of a one-off analysisVisualizationApplications that present meaningful data graphicallyAcquisitionTechniques for cleaning up messy public data sourcesSerializationMethods to convert data structure or object state into a storable format