Archives and Discussion

Please note the following "message archives" are copied from the CC Mailing Archives (members only), put here for easy perusal.
Enjoy!

Titles

• October 4, 2007 -- robots and CodeInvaders talk next Friday (10/12)
• October 4, 2007 -- programming meeting
• October 3, 2007 -- pizza problem
• October 1, 2007 -- Meeting Tomorrow

Full Messages

• robots and CodeInvaders talk next Friday(10/12)

  -- Thu, 4 Oct 2007 11:36 AM
  

  hey all,

  There will be a math/cs talk next week at noon with pizza. Please drop by, especially if you're at all interested in CodeInvaders-- there'll be an example using the game.

  cheers,

  Jonathan

  ---------------------------------

  JONATHAN BREDIN of Colorado College (Rated G)

  /A Coming to Terms:
  Bodies Trump Brains,
  Elephants Can’t Play Chess,
  Boys Blow Things Up,
  and We All Hate Complexity
  (Managing Complexity with a Subsumptive Architecture) /

  Controlling and planning actions for a mobile robot is a classic challenge in artificial intelligence (AI). The traditional approach leverages a sense-model-plan-act loop and found success in formulating and solving complex goals in laboratory settings, but also serves as one of the most visible research directions to succumb to the AI winter—a collapse in AI research and funding lasting into the mid 1980s. Several obstacles stymied the traditional approach. Researchers had trouble modeling real-world settings and logical planning proved fragile in dynamic environments. Furthermore, few projects could scale to utilize multiple processors, accommodate multiple robot goals, or handle conflicting sensor measurements.

  The introduction of subsumption architecture is frequently credited as the paradigm that thawed the AI winter for robotics. The idea coordinates independent modules, each competing to represent a simple low-level behavior, and serves as a useful paradigm to tackle complexity. We’ll look back at a critical moment in AI history, introduce subsumption-style organization, and apply the paradigm through designing a subsumption architecture to control virtual spaceships inside a multi-player programming game. The bulk of the material presented will be accessible to a general audience. No programming experience will be assumed.

• programming meeting

  -- Thu, 4 Oct 2007 11:36 AM
  

  hey all,

  Please feel free to drop by this noon to TSC215 for pizza and programming chats. We'll go over some contest problems and mull around project ideas.

  best wishes,

  Jonathan

• pizza problem

  -- Wed, 3 Oct 2007 12:11 PM
  

  hey all,

  There's a new pizza problem posted on the web: http://www.coloradocollege.edu/dept/MA/

  PIZZA PROBLEM

  Solve the problems posted on the dept bulletin board for a chance to win a free pizza. Problems are posted once a Block.

  Turn your answers into *Math Paraprof, Lauren Bose, in TSC 210. GOOD LUCK!!!*

  Note that you can turn in solutions to unsolved pizza problems. If yours is the first correct solution, then you win a free pizza.

  VIEW OLD PIZZA PROBLEMS
  http://www.coloradocollege.edu/dept/MA/oldpizza.html

  BLOCK 2 2007

  *Computer Science Majors:*

  This is a problem about betting. Each bet works as follows. There are some number of "advisors" and you. One advisor will write either 1 or 0 on a piece of paper, show it to the other advisors but not to you, and put the paper face down in front of you. Each advisor will tell you what the number is. They are all very good actors, so no obvious trick or facial expression will reveal to you who is telling the truth and who is not. The amount you can wager on each bet is anything from $0 to the total amount of money you have.

  You have four bets, but there is a "partial truth teller". That advisor is not guaranteed to tell the truth all the time, but must do so three out of four times. Further, the advisors can actually change the number on the paper once they hear your bet. However, they cannot change the result if doing so would eliminate the possibility that at least one of them is a partial truth telling advisor.

  What can you guarantee to win in four bets, starting with $100, with four advisors, three of whom can lie at will, and one who must tell the truth at least three out of four times?

• Meeting Tomorrow

  -- Mon, 1 Oct 2007 7:31 PM
  

  This is a reminder to everyone that we will be meeting tomorrow at 7pm in the Math Lounge. There should be a few more people this time, so we will do introductions again, then get into the lab to start coding.

  And as a side note, you are more than welcome to bring your laptop.

  Guillermo

Full list of Mailing Archives (members only)