Mohlo by sa vám tiež páčiť

Viac informácií o knihe Edge AI Computing

• Parametre knihy
• Anotácia kniha
• Obľúbené z iného súdka

Nákupom získate 65 bodov

Anotácia knihy

What happens when your device must think for itself-with no cloud, no signal, and barely enough RAM to run a spreadsheet?
The migration of artificial intelligence from data centers to resource-constrained endpoints represents one of the most significant architectural shifts in modern computing. This book addresses the engineering discipline of deploying, optimizing, and maintaining machine learning models on devices where memory is measured in kilobytes, connectivity is intermittent or nonexistent, and every milliwatt counts.
Inside, you will learn:
• How to shrink neural networks by 90% using quantization mathematics and operator fusion, without sacrificing real-world accuracy • Bare-metal deployment patterns for ARM Cortex-M, NPUs, and FPGA fabric that turn milliwatts into meaningful inference • Production-hardened architectures using TensorFlow Lite Micro, ONNX Runtime, and microTVM with deterministic real-time scheduling • Power management protocols including dynamic voltage scaling, sleep orchestration, and energy harvesting for always-on endpoints • Secure over-the-air updates, federated learning loops, and CI/CD pipelines engineered for firmware-grade reliability
Moving beyond theoretical frameworks, this text examines the complete stack: from bare-metal inference engines and heterogeneous compute orchestration to cryptographic storage limitations. You will learn to navigate the constraints that define edge environments-thermal envelopes, deterministic latency requirements, and power budgets-while maintaining model accuracy and production reliability.
If you engineer systems where the cloud is a liability and latency is a physical constraint, this book provides the architectural patterns and optimization techniques to deploy intelligent computation at the true edge.

Parametre knihy

• Celý názov: Edge AI Computing
• Podnázov: TinyML, Embedded Inference & On-Device Optimization
• Autor: Alex Codewell
• Jazyk: Angličtina
• Väzba: Brožovaná
• Počet strán: 318
• EAN: 9798195496890
• ID: 52276715
• Nakladateľ: Independently published
• Hmotnosť: 510 g
• Rozmery: 244 × 170 × 17 mm
• Dátum vydania: 04. May 2026

---

---

---