CS 260: Seminar in Computer Science: Numerical Simulation

General

• Lecture: MWF 3:10-4:00 PM, Gordon Watkins Hall 2141
• Textbook: None

Instructor

Craig Schroeder

Office: Chung 309
Office Hours: MWF 4:00-5:15 PM (after class), or by appointment

Email: craigs@cs.ucr.edu

Course Goals

This course is designed to facilitate students that are interested in learning how to simulate something specific but are not familiar with simulation. It is also assumed that some of the students in the class will be familiar with simulation, and the class is intended to accommodate both. At the beginning of the course, each student will choose a peer-reviewed paper describing a numerical method for simulating what they are interested in. At the end of this course, you should understand your paper and be able to present and explain that paper to the rest of the class. You should also have developed enough background on the topic to read and understand related papers on your own. Students will not be asked to implement anything, and there are no exams in this class.

Course structure

For this course, students will be expected to

• Choose a goal paper. The topic of the paper is up to the student. The difficulty level of the paper should be chosen based on the time frame of the course and the need for other students to be able to follow and contribute to the discussion. It is understood that students will come into the course with different levels of familiarity with mathematics and numerical simulation, and the paper selections should reflect that. Papers with many layers of complexity or lots of moving parts should generally be avoided.
• Choose two additional papers. These two papers are intended to be stepping stones towards understanding the goal paper. If appropriate, these papers might actually be book chapters or from some other appropriate source. They should be simpler and narrower in scope than the goal paper. The first paper should be relatively simple and limited in scope and will likely not be a peer reviewed paper.
• For each of these three papers, students will
  • Study and understand the paper.
  • Schedule a day to present the paper to the class.
  • Make a 20-minute presentation on the paper to the class.
• Attend the other students' presentations and participate in the discussions. Many of the core ideas in simulation are reusable; you may pick up ideas from other papers that help you to better understand you own paper.

Presentations should be 20 minutes. You may use slides or the blackboard/whiteboard for your talk. You will be expected to be able to answer basic questions about your paper during class from students and from the professor. Please practice your presentation before you present. As a general rule, you should understand (1) what problem the numerical method is trying to solve, (2) what the numerical method is and what steps it performs, (3) how/why the method works, and (4) any limitations of the method that are mentioned in the paper. Your presentation should cover each of these four parts, but most of the time should be spent on parts 2 and 3. Part 3 is both the hardest as a reader to understand and the most important for your presentation to cover in detail. It will be very difficult for your audience to understand the method if you do not cover 3 thoroughly. You will not be expected to understand (1) why the method was designed the way that it was, (2) why the paper has the limitations that it does, or (3) whether or why the paper is better or worse than any other method. These are great topics for the class discussion that follows the presentation.

You are advised to start studying your paper at least a week before you are scheduled to present it so that you will have plenty of time to seek assistance if you get stuck. Expect to spend about 4 hours understanding each paper; if you find that it is taking significantly longer than this, your paper may be too hard. Let the professor know so that adjustments can be made. You are permitted (and even encouraged) to read and discuss papers with others in the class, but the papers must be presented individually. Every student must select an unique set of papers.

Schedule

Announcements

• CS 260: Seminar in Computer Science: Numerical Simulation.

Grading