210 likes | 224 Views
Explore computation, complexity, and emergence with Prof. Marie desJardins. Learn about AI, machine learning, and more. Dive into complexity and agents, self-similarity, and sources of complexity. Develop critical analysis skills and experiment with artificial complex systems. Enhance your understanding of iterative, recursive, parallel, and adaptive patterns. Engage through journal entries, lectures, labs, and discussions for active learning.
E N D
Computation, Complexity and Emergence: Course Overview Prof. Marie desJardins, January 30, 2012
Summary of Class • Who am I? • What is “Computation, Complexity, and Emergence?” • What will you learn? • What will you do? • What else do you need to know?
About the Instructor • Prof. Marie desJardins • A.B. Engineering, Harvard 1985 • Ph.D. Computer Science, Berkeley 1992 • Research Scientist, SRI, 1991-2001 • Professor at UMBC 2001-now • Tenured in 2007; promoted to Professor in 2011 • Married since 1985, two daughters (ages 15 (sophomore in high school) and 17 (freshman in college)) • Like to read, do crossword puzzles, sing, play the piano, cook, ski, travel, eat great food
My Research • Artificial intelligence • How to get computers to behave in ways that we would consider to be “intelligent” • Machine learning: Making computers adaptable and able to improve their performance over time • Planning: Enabling computers to make decisions and solve problems • Multi-agent systems: Getting computers to interact effectively with each other and with humans
Research Projects • Integrating low-level learning and planning to enable “skill bootstrapping” (the ability to layer successively more complex skills on top of each other to solve hard problems) • Modeling trust in dynamic real-world environments like supply chain management • Analyzing planning problems to construct libraries of contingency plans for rapid replanning • Preference learning to discover users’ preferences in online rating systems and other applications • Learning from sparse data, and understanding uncertainty in learned models
Complexity and Agents • Complexity in systems arises from interactionsbetween individual components, or agents,of the system • Emergenceis the concept that system behavior is not readily inferred from individual agent behaviors: it arises from the interactions between the agents in complex and beautiful ways • Self-similarity arises when similar patterns occur at multiple levels of abstraction or multiple parts of a system • Sources of complexity: • Parallelism • Recursion • Adaptation
Learning Objectives • By the end of the seminar, students should be able to: • Analyze and explain the ways in which simple individual behaviors and decisions can lead to complex and meaningful global behaviors. • Identify and analyze the sources and effects of complexity in natural and artificial systems. • Demonstrate their understanding of the dynamics of complexity by designing, modifying, and experimenting with artificial complex systems. • Identify, design, and analyze examples of iterative, recursive, parallel, and adaptive patterns in complex systems.
Course Syllabus and Schedule • http://www.cs.umbc.edu/~mariedj/complexity/2012/
Journal Entries • Journal entries are required for each reading assignment listed on the syllabus. • Journal entries must be posted to the course blog no later than 12 hours before class (i.e., by 10pm on Sunday or Tuesday). You are encouraged to post earlier than this, and are welcome to post additional comments in response to other students’ postings. • Journal entries are based on the assigned reading for a given class, and must be posted before that class date. For example, you should post your comments on Chapter 1 by 10am on Sunday, 2/5, since that is the reading assignment for Monday, 2/6. • Journal entries may include questions, comments about the reading, thoughts about other ideas that occurred to you while reading, links to related sources, or just about anything connected to the reading. • Journal entries should be substantive, and should show that you read and thought about the assigned reading, but don’t need to be lengthy. • Journal entries will be graded based on participation portfolios to be submitted by students three times during the semester, as described in the Reading Journal Guidelines handout.
Class Activities • There are three types of class sessions: • Lecture/exercises: I will give a brief lecture/discussion on the topics for the day, followed by some interactive or small-group activities to engage the class in active learning. • Labs: We will have several class sessions in computer labs, to be announced and posted on the schedule in advance. • Discussion sessions: All students are expected to come to class prepared to ask and answer questions related to the reading. (Preparing for class discussion is one of the main purposes of the journal entries.)
Class Participation • All students are expected to participate in class discussions. • Vocal students are expected to make an effort not to dominate the conversation and to give other students a chance to speak. • Quiet students are expected to make an effort to come to class prepared with thoughts they would like to share, and to take the opportunity to share their opinions and insights. • The instructor is expected to ensure that all students have the opportunity to speak and that the discussion is respectful and inclusive!
Groupwork • You may work on assignments in groups, but must write out your own answers (not just copy down a group answer or an answer from the textbook or another source). • Students are always welcome to talk to each other about the course readings, assignments, and concepts. • For written assignments, you should always cite your sources, and quote any material that you wish to include directly. Unquoted material from any source (including the textbook) will be considered an academic integrity violation and dealt with according to the course policy.
Academic Integrity • Please Note: I don’t particularly like policing students, and I hope not to have to worry about this issue at all in an honors course. On the other hand, it is extremely unfair to the other students if some students are not doing the work in the way that it has been assigned. • I take plagiarism and cheating very seriously. You should do your own work, and it is never acceptable to present somebody else’s work as yours. If you get an idea for a project, answer, or commentary from some source, then you should cite that source, just as you would do in a research paper. • If you are using more than two or three words in a row from some source, that is a direct quote; it should be in quotation marks, with a bibliographic citation at the location of the quote. Unquoted text is plagiarism, even if the source is listed in a bibliography.
Getting Help • If you get help from any source (e.g., debugging a program), you should explicitly state that source and the nature of the help. • Reasonable amounts of help (another student helped you to understand what a question was asking, or made suggestions about what tests to run on your program to get it working) are fine. • Excessive help means that you weren’t able to do your work yourself. For example, if somebody else is debugging your programs for you, then it is not your work. • If getting somebody else to debug your program for you is the only way you can get it to work, then it is academically dishonest not to tell me this. If this occurs, you should say so when you submit your assignment, and you should expect to receive a grade that reflects the fact that you did not complete the assignment yourself.
Late Work • All work is due at the beginning of class on the due date, unless otherwise stated. • Late journal entries will receive a 50% penalty, as indicated in the participation portfolio guidelines. • Any work that is to be reviewed by another student (presentations and papers to be peer-reviewed) must be submitted on time in order to receive credit. • Other late assignments will be assessed a penalty as follows: • 0-24 hours late: 25% penalty. • 24-48 hours late: 50% penalty. • 48-72 hours late: 75% penalty. • After 72 hours: No credit. • Extensions of up to one week, in case of circumstances such as travel or other anticipated commitments, may be granted if requested well in advance. Repeated requests for extensions will be denied. • I will grant exceptions to this policy only under the most dire extenuating circumstances, such as a death in the family or incapacitating illness, and only if that situation is documented.
Assignment #1 • Please turn in the student survey after class today. • The “Complexity in Everyday Life” assignment is due next Monday, February 6.