Introduction To Artificial Intelligence CSC 415-515 Where does AI fit in Computer Science? CSC is often a middleman between the computing platform and an application While some problems require simple data processing (e.g., bank or credit card account statements), others require more sophisticated processing arising from a more complex problem. Does computational complexity reflect intelligence or significance of meaning? In some situations, we may may even seek to replace humans in an application and therefore must mimic their performance. For example, one goal of the US Navy in the 21st century is to provide current levels of combat readiness with one-third the crew.
What Is (Artificial) Intelligence? What is intelligence? The ability to learn or understand from experience The ability to acquire and retain knowledge The ability to respond quickly and successfully to a new situation The ability to use reason to solve problems BTW What is knowledge? The act, fact, or state of knowing, including acquaintance or familiarity, awareness, and understanding; Acquaintance with facts; All that has been perceived by the mind; learning. What does it mean “to know”? Have a clear perception or understanding (How do we measure this?); To be aware of, well informed about, or sure of; To have securely in the memory What is reason? An explanation, justification, or motive The ability to think, from judgments, or draw conclusions from known or assumed facts (BTW, “facts” are “the situation as it is”, “reality,” or “the truth”! One might prefer a term such as “evidence” rather than “fact”) CSC 415, Introduction to AI 12/6/2017
What Is Artificial Intelligence? If intelligence is learning, understanding, retaining, responding, and using reason then what is AI? Is mimicking such phenomena using a machine an expression of artificial intelligence? If one finds a clock keeping perfect time, to what would he ascribe intelligence? CSC 415, Introduction to AI 12/6/2017
The Turing Test Place both a human and a machine mimicking human responses outside the field of direct observation and use an unbiased interface to interrogate them. If the responses are distinguishable, the machine is not displaying intelligence. CSC 415, Introduction to AI 12/6/2017
Two Foci of AI Symbol processing Connectionist processing An early thrust of AI research Stereotyped by expert systems, knowledge bases, and inference engines Connectionist processing Typified by emergent computation Often leads to representations possessing large degrees of parallelism Both foci were present in the earliest phases of modern AI research (circa, 1945) and symbolic processing dominated study for several decades. This leadership was in large measure to to the state of technology favoring digital (versus analog) computation However, in the mid-seventies CSC 415, Introduction to AI 12/6/2017
An AI Processing Paradigm Sensor Systems Response Planner Input Stimulus Notice that some processing may be linear (feedforward), while other times closed loop (feedback) processing may be required. Effector Systems Output Response CSC 415, Introduction to AI 12/6/2017
Some Key Contributors: A Note on Origins McCulloch and Pitts (perceptrons) Turing, von Neumann, Shannon, and McCarthy Rosenblatt (perceptron learning) Minsky and Papert Widrow and Hoff (Adaline) Zadeh (fuzzy logic) Werbos, Rumelhart, McClelland, Hinton, Parker, Le Cun Grossberg, Hopfield Holland, Goldberg, De Jong, Koza CSC 415, Introduction to AI 12/6/2017
What Will We Study? Algorithms and Representation Sensor systems Primarily one- and two-dimensional signal sources such as sounds, image data, sunspot counts, or values in a chaotic time series Other sensor data representation such as ensemble codes and transforms Response planning Motion planning—TSP, resource allocation, scheduling Search strategies—Swarms, GAs, Neural networks Effector systems Robot action control CSC 415, Introduction to AI 12/6/2017
Robot Demos Simple line following Find-and-grab Robot tag Robot-to-robot communication (basis for cooperation) Multi-sensor processing Two sensor, line following CSC 415, Introduction to AI 12/6/2017