In Edge Learning, training data are non-independently and identically distributed (non-IID). Applying the same learning strategy for all workers fails to work efficiently. In this chapter, we introduce federated learning, where the training data are always non-IID due to data isolation. Then, we summarize enabling technologies for efficient training with non-IID data. We also propose a reinforcement learning-based method that takes non-IID property and resource-constrain into consideration and adjusts the hyper-parameters to accelerate the loss descent efficiency.

Chapter 3 continues the focus on methodological concerns, outlining and demonstrating how typical acoustic analyses proceeds from fieldwork to acoustic measurements and data extraction to analysis.This treatment goes deeper into the nuts and bolts of "doing" sociophonetics, with more hand’s on procedures for the identification, processing and measurement of vowels and sibilantsThis chapter is intended to introduce readers to and provide the work flow for some basic analyses using common software.The survey of key methods and fundamental terms presented in this and the previous chapter provides a foundation for reading and understanding the approaches used in sociophonetic literature and in exploring the research presented in subsequent chapters.