ICS 643 Spring 2022 Assessment/Grading

The approach to assignments, exams, and relative weighting is intended to assess multiple aspects of your developing expertise in design and analysis of parallel algorithms. In summary, the components and their default weights (percentage of the overall grade) include:

Class Participation: 10%
Homework Problems: 40%
Midterm Exam: 20%
Research paper presentation: 30%

We reserve the right to adjust the total number of points depending on circumstances.

Points, Percents and Letter Grades

To determine letter grades, we use a 4-percent spread per grade increment, i.e., 100-97=A+, 96-93=A, 92-89=A–, 88-85=B+, 84-81=B, 80-77=B–, 76-73=C+, 72-69=C, 68-65=C–, 64-61=D+, 60-57=D, 56-53=D–, 52-0=F. To adjust for variability in the exam difficulty between the semesters, the exams will be graded on a curve with a B average. If upon inspection of the distribution of grades we feel that too many students who understand the material are not getting the grades they deserve, we may then make adjustments in favor of students (especially for those who did well on exams).

Components

Class Participation and in-class quizzes (10%):

The lectures will be fairly interactive: throughout the lecture I will ask questions that will test your understanding of the material being presented. Students are also encouraged to ask questions about anything that is not clear. Student answers and actively asking questions will contribute to the participation grade.

Homework Problems (40%):

There will be 4 to 8 homeworks throughout the semester. You are encouraged to work on the homework problems with other classmates. However, the final write up must be performed individually, in own words, that means copying solution from each other is NOT permitted. Using internet to find solutions to homework problems is also NOT permitted and will be treated as plagiarism (see Class Policies for more details). Exam questions often are similar to homework problems, so this is your chance to make sure that you understand concepts and can work out problems on your own within the time pressures of the exam, not just in a group context.

The homework must be typed up - handwritten homeworks will receive no credit. You should use a good typesetting software that has high-quality equation editor and pseudocode typesetting. I highly recommend Latex and using algpseudocode package. If you use other typesetting software, make sure you use proper indentation in your pseudocode and don't rely on brackets to define your code blocks. Avoid using ASCII symbols when a more natural mathematical symbols exist, e.g., use ≠ intead of !=, or properly typeset exponentiation instead of using ^ symbol. Using such simple changes significantly improves readability of your pseudocode by humans.

Start each problem on a separate page. Any problem that does not start on a separate page will be given 0 points. Write down your name on every page, as the problems will be separated for grading and if a page does not have your name we will not know who wrote the solution.

Homework submission: Homeworks should be submitted in Laulima in PDF format. They will be due before the corresponding class period. Since homework solutions will be discussed during the lecture, no late homeworks will be accepted.

Take-home Midterm Exam (20%):

The midterm exam will cover the first half of the course. The exam will be a take-home exam, open book, open notes, but no internet resources except the reading materials listed on the course website are allowed. The midterm exam must be an individual effort with no collaborations allowed. Just like the homeworks, the exam solutions must be typed up.

Research paper presentation (30%):

By the end of the semester, the students will be expected to read a research paper on the topic of parallel algorithms, understand the presented algorithm, prepare a 30-minute presentation and present the paper to the rest of the class, with the goal of teaching your classmates a new algorithm. You will be graded on the quality of the prepared slides, the clarity of the presented details of the algorithm, and your ability to answer questions about the presented algorithm. The rest of the students will be expected to give feedback on each presentation, which will contribute to their class participation grade.

The list of potential papers to choose from is presented below. However, if you have a paper you'd like to present, talk to me to make sure the paper has enough algorithmic novelty to be appropriate for the class.

ICS 643 Spring 2022 Assessment/Grading

Points, Percents and Letter Grades

Components

Suggested Research Papers for Final Presentation