Bøger af Qiang Yang

How is it possible to allow multiple data owners to collaboratively train and use a shared prediction model while keeping all the local training data private?Traditional machine learning approaches need to combine all data at one location, typically a data center, which may very well violate the laws on user privacy and data confidentiality. Today, many parts of the world demand that technology companies treat user data carefully according to user-privacy laws. The European Union's General Data Protection Regulation (GDPR) is a prime example. In this book, we describe how federated machine learning addresses this problem with novel solutions combining distributed machine learning, cryptography and security, and incentive mechanism design based on economic principles and game theory. We explain different types of privacy-preserving machine learning solutions and their technological backgrounds, and highlight some representative practical use cases. We show how federated learning can become the foundation of next-generation machine learning that caters to technological and societal needs for responsible AI development and application.

What is it like to be a researcher or a scientist? For young people, including graduate students and junior faculty members in universities, how can they identify good ideas for research? How do they conduct solid research to verify and realize their new ideas? How can they formulate their ideas and research results into high-quality articles, and publish them in highly competitive journals and conferences? What are effective ways to supervise graduate students so that they can establish themselves quickly in their research careers? In this book, Ling and Yang answer these questions in a step-by-step manner with specific and concrete examples from their first-hand research experience. Table of Contents: Acknowledgments / Preface / Basics of Research / Goals of Ph.D. Research / Getting Started: Finding New Ideas and Organizing Your Plans / Conducting Solid Research / Writing and Publishing Papers / Misconceptions and Tips for Paper Writing / Writing and Defending a Ph.D. Thesis / Life After Ph.D. / Summary / References / Author Biographies

The great challenge of reverse engineering is recovering design information from legacy code: the concept recovery problem. This monograph describes our research effort in attacking this problem. It discusses our theory of how a constraint-based approach to program plan recognition can efficiently extract design concepts from source code, and it details experiments in concept recovery that support our claims of scalability. Importantly, we present our models and experiments in sufficient detail so that they can be easily replicated. This book is intended for researchers or software developers concerned with reverse engineering or reengineering legacy systems. However, it may also interest those researchers who are interested using plan recognition techniques or constraint-based reasoning. We expect the reader to have a reasonable computer science background (i.e., familiarity with the basics of programming and algorithm analysis), but we do not require familiarity with the fields of reverse engineering or artificial intelligence (AI). To this end, we carefully explain all the AI techniques we use. This book is designed as a reference for advanced undergraduate or graduate seminar courses in software engineering, reverse engineering, or reengineering. It can also serve as a supplementary textbook for software engineering-related courses, such as those on program understanding or design recovery, for AI-related courses, such as those on plan recognition or constraint satisfaction, and for courses that cover both topics, such as those on AI applications to software engineering. ORGANIZATION The book comprises eight chapters.

Indeed, the basic description of a plan generation algorithm has remained constant for nearly three decades: given a desciption of an initial state I, a goal state G, and a set of action types, find a sequence S of instantiated actions such that when S is executed instate I, G is guaranteed as a result.