|
|
||||
|
|
|
|
|
|
Declarative Methods
|
|
| Lectures: | MWF 3-4 pm, Hodson 316 |
| Prof: |
Jason Eisner - 
( )
|
| CA: |
Leah Hanson -
|
| Discussion session: | CA-led session (optional) for activities/discussion/questions/review: TBA |
| Discussion site: | http://piazza.com/class#spring2012/600325 |
| Office hrs: |
For Prof: MW 4pm after class; or by appt in Hackerman 324C
For CA: TBA |
| Web page: | http://cs.jhu.edu/~jason/325 |
| Textbook: | None, as I don't know of any other courses like this one. Suggestions welcome. I may assign some readings. |
| Computer accounts: |
You will need an account on the CS undergrad machines
(ugrad1-ugrad18, etc.). That is where the software will be installed,
and where your work will be tested. It's okay to work on some other machine where you have installed the software yourself -- but you must make sure your code runs on the ugrad machines before you hand it in. |
| Policies: |
Academic integrity: read this!
It says your work must be your own, etc. Homework submission procedure: TBA Lateness: floating late days policy Announcements: Read mailing list and this page! |
| Requirements and grading: |
For 600.325 students: 5% class participation (this makes a difference!), 50% assignments, 15% midterm, 30% final. For 600.425 students: 25% (TBA) for term project, with the rest partitioned as for the 325 students. You might also be assigned extra reading, extra questions, or a class presentation. |
600.325/425. Declarative Methods (3 credits). Suppose you could simply write down a description of your problem, and let the computer figure out how to solve it. What notation could you use? What strategy should the computer then use? In this survey class, you'll learn to recognize when your problem is an instance of satisfiability, constraint programming, logic programming, dynamic programming, or mathematical programming (e.g., integer linear programming). For each of these related paradigms, you'll learn to reformulate hard problems in the required notation and apply off-the-shelf software that can solve any problem in that notation -- including NP-complete problems and many of the problems you'll see in other courses and in the real world. You'll also gain some understanding of the general-purpose algorithms that power the software. [Analysis]
Prereq: 600.226, Calc II. Students can only receive credit for 600.325 or 600.425, not both.
You could regard this as an alternative programming course. A programming course teaches you how to use a programming language to solve problems. It also outlines how your computer will actually execute the code you write in that programming language.
The languages we'll be using aren't conventional programming languages. They are powerful problem description languages that focus on special-purpose computation. These languages are declarative. That is, you use them to specify a problem, not a solution. But they are backed up by solvers that do a good job of finding the solution efficiently in most cases. This course will survey some declarative languages and examine the kinds of solvers that people have written for them.
How do these languages relate to conventional ones? Conventional languages help you build arbitrary large systems in a modular way. You can think of a solver as a particular powerful module that handles many problems of a particular sort. To explain your particular problem to the solver, you use a declarative language.
This class is in the 3-4:30 time slot, a "flex slot" that is intended to permit either three 50-minute lectures or two 75-minute lectures per week.
Usually we have three 50-minute lectures per week, on MWF. However, please keep the entire 3-4:30 timeslot open, especially on MW:
A typical unit will last two weeks. The first week will cover a particular declarative language. The second week will look at strategies used by modern solvers for that language. Your homework (assigned at the end of the first week) will ask you to program in the language, using specific solver software.
Warning: Schedule below is subject to change.
Warning: For future lectures and assignments, the links below take you to last year's versions, which are subject to change.
Here are some materials that we covered in past years but will probably omit this year.
(This topic is now covered well in 600.475 Machine Learning.)