Conflict of Interest
CMT has a very conservative conflict of interest management system. As a result, it may be helpful to define conflicts of interest.
A reviewer cannot review papers, and area chairs cannot manage papers, authored by:
themselves;
their current or former students;
their current or former post-doctoral fellows;
their current or former thesis advisors;
their current or former thesis post-doctoral fellows;
co-authors on papers in the previous 3 years;
the principal investigator on a currently active grant for which they are a co-investigator;
a co-investigator on a currently active grant for which they are the principal investigator;
faculty, students or researchers at their current institution;
close collaborators.
The program chair cannot be an author on any papers, or on any papers that he or she contributed materially.
Note that co-investigators on the same grant are not considered to be in conflict unless they are also close collaborators. There are a large number of multi-PI grants in the US and Europe that fund many roboticists who do not otherwise interact. Disallowing reviewers or area chairs from reviewing or managing papers from co-investigators on a 40- or 50-PI grant would lead to unmanageable conflict of interest situations.
Note that CMT prevents a reviewer from being assigned to a paper if they have a conflict with the same institution as an author on the paper. This is clearly incorrect. I (Nick Roy) have close collaboration with faculty at CMU in the last 3 years. According to the CMT instructions, I should list cmu.edu as a conflict domain. However, imagine that Frank Dellaert, who also has close collaboration with faculty at CMU, also lists cmu.edu. Because the conflicts lack any directionality, CMT will then conclude that Frank and I are in conflict, even though we do not meet any of the conditions above.
If a reviewer is assigned to review a paper for any of the above conditions, we would ask the reviewer to notify the area chair and recuse themselves.
If an area chair is assigned to manage a paper for any of the above conditions, we would ask the area chair to notify the program chair and recuse themselves.
During the area chair meeting to make paper decisions, area chairs will be asked to leave the room for any papers for which they have a conflict of interest. All other area chairs will be welcome to participate in the discussion. Area chairs will not have access to all the papers until the day of the meeting (and will not have access to reviewer identities for any papers they did not manage), so in some cases, the conflict of interest between area chair and authors on the paper will not be noticed until the discussion is to begin. To minimize this risk, we will begin the discussion of each paper by checking for unnoticed conflicts of interest.