Artificial Intelligence (AI) intends to automate, with computer programs and devices, routine and intellectual tasks that heretofore required significant human intervention. In fact, any task, that requires "humanly" intelligence to accomplish, can be considered a targeted application for AI. It encompasses many diverse fields of science, but can be broadly grouped into four categories: acting humanly, acting rationally, thinking humanly and thinking rationally (Artificial Intelligence, 2nd Edition, 2003, by Russell and Norvig). The first two are concerned with behavior, and the latter two address the thinking process. Acting with human-like intelligence may be the most classic AI application, entailing subjects such as natural language processing, knowledge representation, machine learning, computer vision and hearing, and robotics. Acting rationally includes mainly rational agents and rational decision making. Thinking humanly tries to understand and mimic the human cognition process using computers, which are also the objectives of cognitive science. The General Problem Solver (GPS) project was the first attempt in this direction. Thinking rationally seeks to formalize in computing models the "right thinking" (not humanly thinking), with most of its principles derived from the study of Logic. Many computational reasoning systems belong to this category. To gain a more rigorous theoretical foundation, AI has incorporated into it many other scientific disciplines, such as statistical modeling, probability, belief network, decision theory, game theory, fuzzy logic and operational research. All of these are now considered part of AI. Some specific applications from decades of AI development include planning and scheduling, game playing, autonomous control, diagnosis systems, logistics planning, language understanding, contextual information retrieval and robotics.