How to Assign Papers to Referees
The problem to assign papers to referees gained a considerable interest in the recent years, especially in the scope of conference management systems.
TU Wien, Campus Favoritenstraße
Seminarraum FAV EG C Gödel
1040 Vienna, Favoritenstraße 9-11
Erdgeschoß, HBEG10, HBEG10
The Database and Artificial Intelligence Group of Institute of Information Systems and the Theoretical Informatic and Logic of the Institute of Computer languages invite to the following talk:
The problem to assign papers to referees gained a considerable interest in the recent years, especially in the scope of conference management systems. These systems need to achieve a fair and balanced distribution of papers among referees, where the conditions of fairness and balance may be defined in several ways. We present two algorithms to distribute a possibly large number of papers among a smaller number of referees, each paper requiring k reports. The first one is an approximation algorithm, using an iteration of weighted maximum matching in bipartite graphs. The second one is an exact algorithm based on b-matching. The optimality criterion for the assignment is not based on a local view of each referee, but on a global performance of the whole k-assignment satisfying a fairness criterion. We introduce an objective function for the k-assignment problem ensuring a specific notion of fairness when it is maximized. We show how a few precisely defined fairness criteria can be achieved that way. This includes a particularly notable extension of rank-maximality, a notion introduced by Irving et al. Our second algorithm computes in polynomial time optimal k-assignments with respect to the aforementioned fairness function.
With kind support of the Wolfgang Pauli Institute
- Miki Hermann, LIX, Ecole Polytechnique Palaseau, France
Note: This is one of the thousands of items we imported from the old website. We’re in the process of reviewing each and every one, but if you notice something strange about this particular one, please let us know. — Thanks!