• Combines historic document analysis and empirical micro-level quantitative data. • The research is comprehensive, focusing on both urban and rural areas in China. • Wage's negative effect on the teaching profession is less discussed in the academic field. • The first volume to address teacher occupational choice in China.

* Combines historic document analysis and empirical micro-level quantitative data. * The research is comprehensive, focusing on both urban and rural areas in China. * Wage's negative effect on the teaching profession is less discussed in the academic field. * The first volume to address teacher occupational choice in China.

Deep Learning on Edge Computing Devices: Design Challenges of Algorithm and Architecture focuses on hardware architecture and embedded deep learning, including neural networks. The title helps researchers maximize the performance of Edge-deep learning models for mobile computing and other applications by presenting neural network algorithms and hardware design optimization approaches for Edge-deep learning. Applications are introduced in each section, and a comprehensive example, smart surveillance cameras, is presented at the end of the book, integrating innovation in both algorithm and hardware architecture. Structured into three parts, the book covers core concepts, theories and algorithms and architecture optimization. This book provides a solution for researchers looking to maximize the performance of deep learning models on Edge-computing devices through algorithm-hardware co-design.

Workflows may be defined as abstractions used to model the coherent flow of activities in the context of an in silico scientific experiment. They are employed in many domains of science such as bioinformatics, astronomy, and engineering. Such workflows usually present a considerable number of activities and activations (i.e., tasks associated with activities) and may need a long time for execution. Due to the continuous need to store and process data efficiently (making them data-intensive workflows), high-performance computing environments allied to parallelization techniques are used to run these workflows. At the beginning of the 2010s, cloud technologies emerged as a promising environment to run scientific workflows. By using clouds, scientists have expanded beyond single parallel computers to hundreds or even thousands of virtual machines. More recently, Data-Intensive Scalable Computing (DISC) frameworks (e.g., Apache Spark and Hadoop) and environments emerged and are being used to execute data-intensive workflows. DISC environments are composed of processors and disks in large-commodity computing clusters connected using high-speed communications switches and networks. The main advantage of DISC frameworks is that they support and grant efficient in-memory data management for large-scale applications, such as data-intensive workflows. However, the execution of workflows in cloud and DISC environments raise many challenges such as scheduling workflow activities and activations, managing produced data, collecting provenance data, etc. Several existing approaches deal with the challenges mentioned earlier. This way, there is a real need for understanding how to manage these workflows and various big data platforms that have been developed and introduced. As such, this book can help researchers understand how linking workflow management with Data-Intensive Scalable Computing can help in understanding and analyzing scientific big data. In this book, we aim to identify and distill the body of work on workflow management in clouds and DISC environments. We start by discussing the basic principles of data-intensive scientific workflows. Next, we present two workflows that are executed in a single site and multi-site clouds taking advantage of provenance. Afterward, we go towards workflow management in DISC environments, and we present, in detail, solutions that enable the optimized execution of the workflow using frameworks such as Apache Spark and its extensions.

The year 2009 marks the 30th anniversary of normalization of Sino-U.S. relations. Over the past 30 years, the bilateral relations have developed by twists and turns. It is not until recent years that some stability and forward-looking exchanges have returned to the central stage, albeit tension, grievances, and mistrust continue to persist. Washington has encouraged China to become a "responsible stakeholder" in the world affairs, while China has urged the U.S. to work with China to build a "harmonious world." Both sides want to work together to solve their differences through dialogs and negotiations. In the wake of the worldwide financial crisis of 2008-2009, China has contributed greatly in financing the crumbling U.S. financial market and lent a helping hand in stabilizing the world economy. Nevertheless, the foundation of the relationship remains very fragile and the long-term prospect for a constructive cooperative relationship is still full of uncertainties. For many Americans, China's increasing global reach and growing political and economic influence constitute the greatest challenge to world dominance by the United States. As a result, some perceive China's rise as a threat to Americans' core national interests. The recent changes in the global geostrategic landscape and economic interdependence have suggested that some new ideas, factors, conditions, and elements are shaping the relations between the two countries. The task of Thirty Years of China-U.S. Relations: Analytical Approaches and Contemporary Issues is to explore these factors, issues, and challenges and their impact for the bilateral relations in the 21st century.

Scalable and efficient distributed learning is one of the main driving forces behind the recent rapid advancement of machine learning and artificial intelligence. One prominent feature of this development is that recent progress has been made by researchers in two communities: (1) the system community such as database, data management, and distributed systems, and (2) the machine learning and mathematical optimization community. The interaction and knowledge sharing between these two communities has led to the rapid development of new distributed learning systems and theory. This monograph provides a brief introduction to three distributed learning techniques that have recently been developed: lossy communication compression, asynchronous communication, and decentralized communication. These have significant impact on the work in both the system and machine learning and mathematical optimization communities but to fully realize the potential, it is essential they understand the whole picture. This monograph provides the bridge between the two communities. The simplified introduction to the essential aspects of each community enables researchers to gain insights into the factors influencing both.The monograph provides students and researchers the groundwork for developing faster and better research results in this dynamic area of research.