Docstoc

CSCI 446Artificial Intelligence CSCI 555Applications of Artificial

Document Sample
CSCI 446Artificial Intelligence CSCI 555Applications of Artificial Powered By Docstoc
					              CSCI 446: Artificial Intelligence
       CSCI 555: Applications of Artificial Intelligence
                                          Fall 2012
                              CSCI 446 section 1 (CRN: 75317)
                              CSCI 555 section 1 (CRN: 75318)
                                  TR 11:10 – 12:30 SS 362
Instructor: Alden Wright, 407 Social Science, 4790, alden.wright@umontana.edu
Prerequisites: The official prerequisite is CSCI 205. However, M 225 Discrete Math (or M 207)
    and STAT 341 Intro to Probability and Statistics are much more important.
Office hours: (tentative) Monday, Wednesday 3:10-4:00, Tuesday, Thursday 12:45-2:00.
Class Format: Last fall, Stanford AI professors Sebastian Thrun and Peter Norvig offered a free
    online version of their Stanford AI class. With no promotion, 160,000 students signed up.
    Of these, 39,000 completed the class and 22,000 passed. This course was truly revolutionary
    in that a high quality college course that included graded homework and exams from a
    prestigious university was offered free to such a large number of students. Now two startup
    companies, Udacity and Coursera, are offering similar courses, and Harvard and MIT have a
    joint initiative called EdEx to develop open source software to offer similar courses.
    I have decided to use the online lectures for the Stanford AI class as a "textbook substitute".
    These lectures are at https://www.ai-class.com/home/. These lectures include integrated
    feedback quizzes which are very helpful in understanding the material. My reservation
    about doing this is that the website hosting the course has been erratic. The lectures without
    the integrated feedback are available on YouTube, so there is a partial substitute. If there
    are too many problems with the website, we may need to consider another option.
    In class, we will discus the topics presented in lectures and work problems. Graduate
    students may also present their projects. Since this is an experimental approach, I most
    definitely want feedback on how to improve our use of class time.
    One of the strong points of this approach is that the course will cover up-to-date topics as
    presented by leaders in the field of AI. Sebastian Thrun was the leader of the Stanford
    driverless vehicle team that won the 2007 DARPA challenge and was second in the 2009
    challenge. He is (or was) head of the Google driverless car initiative. Peter Norvig is co-
    author of the leading AI textbook and is head of AI research at Google. One of the major
    recent developments in the field of AI is the incorporation of probability models both in
    knowledge representation and machine learning, and this course will have a heavy emphasis
    on these models. See http://www.nytimes.com/2008/05/03/technology/03koller.html?_r=1
    for a popular article on how another Stanford AI professor has been a pioneer in this area.
    I will be seeking frequent feedback on how this class format is working. If there are
    problems, I will be ready to try something else, including a more traditional lecture-
    discussion format.
Textbook: As of now, there is no required textbook. If you want to buy a textbook, the one to
   buy is "Artificial Intelligence, a Modern Approach" (3rd edition) by Stuart Russell and Peter
   Norvig. A copy of this book is on 2 hour reserve at the library associated with the classes
   CSCI 446 and CSCI 555. The sections of the Stanford AI class taught by Prof. Norvig
    correspond closely to chapters in this book while the sections taught by Prof. Thrun at best
    correspond loosely to this book.
Moodle: There is a Moodle class supplement for this course. (There is one supplement under
   CSCI 446 for both courses.) Assignments and other supplementary material will be posted
   in Moodle, and homework may be turned in via Moodle.
Required Work: There will be homework assignments given approximately once per week.
   Some of the problems will be fairly routine and all students will be expected to complete
   these by the due date for the assignment. Other problems will be more challenging. These
   will have more flexible due dates, and students who do these problems will often be
   expected to present their solutions in class.
    There will also be programming assignments. You can choose the language in which you do
    the assignments. I must be able to compile and run your solutions, and some assignments
    will require a language with classes and objects. I will not cover Lisp in this class.
    Graduate students are required to do an additional project, which is described at
    http://www.cs.umt.edu/~wright/446/ Graduate_Student_Project_Artificial_Intellignece.pdf.

    There will be a midterm and a final. I will try to set the date for the midterm early in the
    semester---remind me.
Grading: The relative contribution of the various assignments, exams, and projects to your final
   grade is yet to be determined. Tentatively, class attendance will not count towards your
   grade.
Incompletes and late drops: I will strictly follow University policy. In for me to consider an
    incomplete or late drop or change to audit status, you will have to submit documentation
    (such as a note from a doctor) to verify your reason for the incomplete or late drop. The
    acceptable reasons for a late drop are limited to: registration errors, accident or illness,
    family emergency, and change in employment schedule. See pages 20 and 21 of the catalog
    for the University policies.
Disabilities: Only disabilities that are certified by DSS will be considered for special treatment.
    If a student has such a disability, I expect to be notified in the first week of class.
Collaboration: Collaboration on problems is OK if you acknowledge and describe the
    collaboration when you turn in the writeup for the problem. If you collaborate and do not
    acknowledge/describe this collaboration, you are cheating, and you may fail the course. You
    are responsible for writing/typing your own assignments, and you are responsible for
    understanding what you turn in. In the case of acknowledged collaboration, I reserve the
    right to ask the student to explain what he/she has done, and to adjust the grade assigned on
    the basis of this explanation. Collaboration on exams is not acceptable.
Topics Covered: See the Stanford AI class at: https://www.ai-class.com/home/. We will
   probably not have time to cover all of the topics in the Stanford AI class. We will have an
   additional section on the social and ethical implications of AI.

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:0
posted:4/12/2013
language:English
pages:2