**Artificial Intelligence**

**Study Guide #3**

**Chapter 11: Planning**

Know the definition of a complete plan. Understand how preconditions are achieved. Be able to discuss the differences between searching and planning. Know the structure of the language of planning problems (representation of states, goals, actions, plan). Know the syntax for STRIPS and CLIPS. Understand the examples in the textbook and notes (airport cargo, spare tire, blocks world). Know the concepts of clobbering as related to partially ordered plans. Understand CLIPS examples passed out in class.

**LISP**

Know the basic concepts of LISP including, but not limited to: t, nil, atoms, lists, s-expressions, arithmetic functions, comparison and boolean functions, cons, car, cdr, setq, LISP data structures and cond. Be able to define (create) your own (simple) user-defined functions in LISP.

**Chapter 20: Learning and Neural Networks**

Know the definition of the following terms: neuron, artificial neuron, neural network, artificial neural network, supervised learning, reinforced learning, unsupervised learning, training data, test data, epoch, decision boundary, (non) linear separability. Know what the threshold logic unit is and the functions that model it: identity, binary step, binary sigmoid, bipolar sigmoid. Know and understand the McCulloch-Pitts neuron model. Know the network structure for the Perceptron (two-layer). Be able to compute the output response for a given network with provided weights. Know and understand the McCulloch-Pitts ANNs for simple logic gates (as presented in the notes). Know the Perceptron learning algorithm and the Perceptron weight update rule.

**Chapter 24: Computer Vision**

Know the two approaches for perception: feature extraction and model based. Know the fundamental steps in digital image processing. Understand that an image is a 2D array of intensity values at a given location at a given time. Understand the concept of Fourier Domain filtering (low, high, band, notch filters). Know and understand the convolution algorithm for filtering and edge detection.

Terms: pixel, image, filter, segmentation, connected component, medial axis transformation. Given a shape, be able to derive the chain code, difference, and shape number of the object. Given an image, be able to compute the Euler number for (non) polygonal shapes.