Machine Learning for Biomedicine

Spring 2009

Instructors: Subramani Mani and Hua XU


T/TH 4:00 pm 5:15 pm Room 456 Eskind Library

Class Policies

Grading: The factors that affect grading and their relative weights are as follows:

1)      Attendance and class participation, 10%

2)      Written assignments and term project, 35%

3)      Midterm Exam, 15%

4)      Final Exam, 20%

5)      Other class activities such as quizzes and paper presentations, 20%


The final grade is the weighted average of the above factors subject to the constraint that the student has satisfactory performance in all the above categories, specifically that he or she has more than 50% of the grade in each category. For example, a student that gets less than 50% in the final exam does not get a passing grade even if he or she has perfect scores in all the rest categories. The activities in case (5) above are to be decided according to class participation, interests, strengths and weaknesses.


Grading of Assignments: No credit will be given for delayed assignments unless a serious condition, for example a serious illness, prevented the student from working on it and the student has the agreement of the instructors to turn in the assignment late. Note that an assignment that was not returned on time will also be graded so that feedback is given to the student.


Grading: The weighted average of performance in the above five categories will provide a score that will be used for the final letter grading as follows:

absolute weighted score >= 92% A+

absolute weighted score >= 88% <92% A

absolute weighted score >= 85% <88% A-

absolute weighted score >= 80% <85% B+

absolute weighted score >= 75% <80% B

absolute weighted score >= 72% <75% B-

absolute weighted score >= 70% <72% C

absolute weighted score >= 65% <70% D

absolute weighted score <65% F


The grade is not negotiable. However students are welcome to point out to the instructor(s) possible errors that have been made in grading.




Make up Exams : A make up exam will be given to any student that was missing during the mid term or final exams, provided that he or she provides proof of a genuine reason for missing the exam, such as a letter from a doctor that certifies an illness.


Asking Questions to the Instructor:


Please note that students should send in advance an e-mail message stating the question to the instructor responsible for the corresponding lecture and/or assignment that generates the question This is so that the instructor can respond to all students if pertinent and to allow them to prepare a suitable reply.


If an issue is not resolved by email, the students can set up an appointment with the instructors.


Format and Rules for Paper Discussions:

1. Students are required to have studied the papers before the discussion day.

2. One person presents a ~20 minute outline of the main topics (goal, methods, results) of a paper.

3. The presenter will typically be a student so they get experience in abstracting and presenting papers.

4. Presenters will alternate/rotate among the students so that at the end everyone will have presented the same number of papers.

5. The presenter need not prepare a highly-polished presentation. Handwritten transparencies and transparency photocopies of tables and graphs from the original are acceptable as means to save preparation time.

6. After the presentation there will be up to 20 minutes of discussion per paper and a occasionally a short quiz (10 minutes or so). The purpose of the quiz (if any) is to make sure students have read the paper carefully ahead of time.

7. The presenter for a paper is excused from the quiz.

8. In general students will assign presentation dates/papers among themselves and only when they cannot agree they will ask the instructor to assign papers.


Honor Code related policy: Unless an assignment is designated as "collaborative" students are not supposed to work together to solve it. They may discuss what is that the assignment problem asks, but not how to do it. If in doubt that such discussions are in violation of the honor code :

they should refrain from discussing them and ask the instructor. In general students should make an effort to ask for clarifications from the instructor unless an assignment is collaborative.


List of important concepts: Students should identify appropriate sources. 95% of the time these sources will be the class textbooks or distributed material. For a small number, (e.g., formal definition of a tree) any acceptable source (e.g., an algorithms book) will suffice. If in doubt about using the appropriate source students should ask the instructor.