Generalizable Commonsense Reasoning for Hybrid Intelligence Systems

This book enables readers to understand the challenges and opportunities of developing trustworthy AI with commonsense reasoning skills. Commonsense knowledge is often implicit and presents a challenge for automated methods in natural language processing and question answering as the extraction and learning algorithms cannot count on the commonsense knowledge being available directly in text. As such, commonsense knowledge and reasoning has been considered the “black matter” of AI, raising concerns about the trustworthiness and applicability of AI methods in automated and hybrid applications, especially social good applications in misinformation, traffic, health, and education. This book presents dominant methods that combine neural and symbolic advances to achieve adaptivity, collaboration, explainability, and responsibility through commonsense reasoning. In addition, the book describes how these socio-technical properties of AI can facilitate a range of social-good applications like misinformation, traffic, education, and health. What makes commonsense reasoning such a unique and impactful challenge? What do cognitive and AI perspectives bring to the table? How can we approach building responsible, adaptive, collaborative, and explainable AI with common sense? And finally, what is the impact of this work on hybrid human-AI intelligent systems? This book provides an accessible introduction and exploration of these topics. In addition, this book: - Features broad coverage of cognitive, symbolic, and neural theories for common sense reasoning - Integrates neural and symbolic methods for open-domain tasks that require adaptivity, collaboration, and responsibility - Synthesizes lessons learned and provides open challenges that future research and engineering efforts need to tackle