دانلود کتاب هوش اجتماعی مرز جدید ادغام هوش انسانی و هوش مصنوعی در فضای اجتماعی

The rise of Artificial Intelligence (AI) and its tight interactions with humans and our society leads to the emergence of a new field called Social Intelligence (SI). The SI is motivated by the complementary strengths of AI and humans. For example, AI is observed to excel at tasks that require high speed, run at a large scale, and generate accurate quantitative results. In contrast, humans often outperform machines at tasks that require critical thinking, creative works, and excellent social skills. Different from many existing works that take AI and humans as competitors in a zero-sum game, this book presents an exciting vision of SI that allows humans and AI to collaborate with each other and build a novel paradigm of collective and hybrid intelligence by fully exploring their complementary strengths and interactions in the social space. The SI will empower human-centered AI solutions in many critical application domains by fully unleashing the power of integrating human intelligence with AI. Examples of such application domains include truth discovery and explanation, healthcare analytics, disaster response, online education, face recognition, intelligent transportation, urban sensing, and smart cities. The SI paradigm introduces a set of critical challenges for research. Examples include human-centered data heterogeneity and sparsity, model generality and adaptability, explainability, fairness and bias, privacy, and hybrid intelligence integration. This book addresses these challenges by presenting a series of principled analytical frameworks and real-world system designs that fully explore the collective strengths of AI and human intelligence, while explicitly addressing the unique concerns and constraints of humans. The book first presents a set of novel human-centered AI solutions (e.g., multimodal approaches, robust and generalizable models, socially empowered explainable AI designs) to address the aforementioned SI challenges. The book then presents several human-AI collaborative learning frameworks that jointly integrate the strengths of collective human intelligence from people and AI to address the limitations of human-only or AI-only solutions. Finally, the book discusses the pressing societal and human-centered issues in SI such as fairness, bias, and privacy. The book also offers extensive evaluation of the discussed SI systems in real-world applications and case studies to demonstrate the effectiveness and performance gains of the presented solutions in comparison to state-of-the-art baselines in different aspects such as model accuracy, generalizability, explainability, algorithmic fairness, and system robustness. Leveraging the models, techniques, and systems presented in this book, the reader is offered with analytical foundations, optimized frameworks, and system prototypes needed to explore the power of social intelligence. The SI paradigm generalizes the current trends of human-AI interactions, human-assisted AI, and AI for social good into a holistic human-AI ecosystem with social context. The book takes the reader on a journey of discovery through the analytical and systematic underpinning of developing novel theories, models, and systems in the domain of social intelligence. The uniqueness of human-centered nature and integration of human intelligence and AI makes this journey more exciting and challenging. The authors hope that techniques developed in this book will become part of the solution space in dealing with challenges in future social intelligence systems. These techniques can help fully unleash the power of both AI and humans in the next generation of computing, intelligence, and information systems.