One of the most difficult problems facing researchers in artificial intelligence has been the inability of traditional models to capture the flexible and robust nature of commonsense reasoning. Ron Sun's innovative work takes a refreshing new approach to this persistent problem by explaining and modeling commonsense reasoning with a combination of rules and similarities, all under a connectionist rubric. Integrating Rules and Connectionism for Robust Commonsense Reasoning surveys areas of rule-based reasoning, connectionist models, inheritance, causality and similarity-based reasoning, and goes on to introduce a new framework and a novel connectionist architecture for modeling commonsense reasoning that synthesizes many of these areas. Along with this framework, the book proposes a set of interrelated new ideas regarding the modeling of commonsense reasoning which are highly relevant to current research in AI and cognitive science and the ongoing methodological debate. The book first analyzes relevant reasoning data and examples to provide insight into this new approach. The analysis establishes a framework for modeling such data based on the notion of rules and similarities, which is then, through detailed experiments and derivations, translated into a connectionist architecture. Because similarity-based reasoning is inherent in connectionist models, and rule-based reasoning is integrated into connectionist networks via encoding rules, the resulting architecture proves naturally capable of carrying out both rule-based reasoning and similarity-based reasoning. The book encourages readers to tap the synergy resulting from the interaction of these two different types of representation and processing. By doing so, they will be able to handle a large number of difficult issues in commonsense reasoning, all within one integrated framework.