CMSI 261 Preparation for the Final ========================= The final exam will be held Thursday, May 3, 2012 at 8 a.m. Yes, 8 A.M. It will be open-book, open-notes, open-computer. No chatting, facebooking, talking, or soliciting answers in any way, of course. The material for the questions will be drawn from: * GEB: Ch 1-14 and 17. * Baum: Ch 1-4, 13, 14. * The Handouts. * The on-line course notes. (There will be some questions involving the use of logical notation, so reading Priest's book might be helpful, but not required. You can get by with the on-line notes on logic.) Expect 20 multiple-choice questions. Here is a brief big-picture review of the course, in order to help direct your study: GOAL OF COURSE: To gain an understanding into the nature and mechanisms of human thought, using tools and methods of computer science. QUESTIONS YOU SHOULD BE ABLE TO ANSWER AFTER TAKING THE COURSE: * What are thoughts, and how are they represented in the physical system we call the brain? * How can we model thoughts (or concepts in general) using language (both formal and natural)? * What are the primary arguments for and against dualism? * Is it likely that one can build machines that are conscious or sentient? HIGH-LEVEL TOPICS: The study of thought requires touching on the following subjects, each of which you should be at least moderately familiar with. You can read the Wikipedia or Stanford Encyclopedia of Philosophy articles on each. * consciousness * understanding * self-awareness * sentience * language * communication * mind * brain * epiphenomena * thought * computation * formal system * logic * meaning * representation * isomorphism UNDERSTANDING GEB: GEB is Hofstader's attempt to show how animate beings can come from inanimate matter, and how strange loops are the key to unraveling the mysteries of consciousness. There are 20 chapters, each preceded by a dialogue. Perhaps you can say there are 21 chapters, since there is an unnumbered chapter which opens the book. This chapter is remarkably "preceded" by the dialogue that appears at the end of the book, after Chapter 20. Pretty cool. Here are what the chapters cover: 1 (MU-puzzle) - What is a formal system? The difference between M and I modes. 2 (Meaning and Form) - "Meaningless" symbols in a formal system acquire meaning simply because of the rules. Meanings have to do with isomorphisms. 3 (Figure and Ground) - It is suggested that it may not be possible to generate the nontheorems of a formal system with a formal system. 4 (Consistency and Completeness) - Consistency and completeness are defined. Consistency is shown to be relative to an interpretation. The notion of "undefined" or "primitive" terms is slippery and may not really even exist. 5 (Recursion) - Recursion exists in nature. It is the way that unbounded complexity arises from a compact description. Without recursion you are limited to the finite. There is recursion in art, music, language, math, the universe, etc. 6 (Location of Meaning) - Meaning may not just be in the message, but also in the player and receiver. There are inner, outer, and frame messages. 7 (Propositional Calculus) - In the PC, "and," "or," "not," and "if-then" are captured in a formal system. 8 (TNT) - The theory of natural numbers is captured in a formal system. 9 (Mumon and Gödel) - Zen and Gödel numbering. 10 (Levels of Description) - The importance of levels within complex systems is discussed and how levels generally only talk to the level below and provide service to the level above. Distant levels are normally shielded from each other. 11 (Brains and Thoughts) - Looks at brain structures to make a pass at the question of how neural activity gives rise to thought. 12 (Minds and Thoughts) - Discusses communication between minds, again using notions of mappings and isomorphisms. Examples are taken from language translation, and the rather interesting "ASU." 13 (BlooP, FlooP, and GlooP) - The big difference between computations specified with only bounded loops and those that use unbounded loops is discussed. Cantor's diagonalization is used to show limits on the power of computation. 14 (Gödel's Incompleteness Theorems) - Introduces use and mention and the quining technique. Gödel's first theorem is proved. 15 (Jumping out of the System) - Gödel's result is shown to apply infinitely to all systems attempting to "repair" themselves to avoid the "problem." Lucas's arguments are criticized. 16 (Self-Ref and Self-Rep) - Covers self-reference in more detail, but also self-replication. A system called typogenetics is introduced to show how some systems can reproduce themselves. 17 (Church, Turing, Tarski, ...) - Here Hofstadter looks at people or systems that seem to "do math" or other tasks "magically." 18 (AI Retrospects) - Programs such as SHRDLU, which existed at the time GEB was written are discussed and it is considered whether they are "thinking" or whether they actually "understand" anything. The Turing Test is also covered here, as is a brief history of AI. 19 (AI Prospects) - Higher-level AI techniques like frames are discussed, as is creativity. Hofstadter gives some of his predictions on the future of AI. 20 (Strange Loops) - Many examples of strange loops. Free will. Consciousness. OTHER MATERIAL: As you go over the handouts, lecture notes, and additional readings, make sure you are comfortable with: * What Dennett and Pinker have to say about consciousness * What Eric Baum says that thought really is, and how and why it evolved the way it did * What Searle's beefs are with strong AI THINGS YOU SHOULD KNOW FOR THE EXAM: Computer Science, and especially AI, deals a lot with knowledge representation, so much of the final will be concerned with formal systems, logic, linguistic notations, and so on. Because these issues take a great deal of practice to master, the exam will be multiple-choice. It is easier to recognize a correct translation into logical notation than it is to generate one on your own. At least that is the hypothesis that led me to the multiple-choice idea. Here are things you should know that would be helpful: * How statements are represented in the propositional calculus, or TNT, or any other logistic system. * How to generate sentences from the grammar given in the notes on recursion. * What Gödel's (first incompleteness) theorem does and does not say. * What a strange loop really is. * How to detect pieces of an isomorphism. * How to detect figure and ground. * How to detect the self-similar copies in a recursive figure. * How to detect the inner, outer, and frame messages. * What a quine is. * How to make a self-referential statement with a quine. * How to identify a paradoxical statement * Exactly what is meant by soundness, completeness, consistency, and decidability. * How the Contracrostipunctus relates to the limits of formal systems. * How Zen strings relate to metalogistic concepts. * How the Magnificrab (Ch. 17 dialogue) is ... impossible. * How people have refuted Lucas (easily). * Searle's main arguments, and what Baum thinks of them. (Wikipedia may help.)