Architecture of a Conference Management System Providing Advanced Paper Assignment Features

TL;DR

This paper presents an advanced conference management system architecture that automates reviewer-paper assignment using semantic keyword hierarchies, reviewer preferences, and conflict detection to improve accuracy and balance.

Contribution

It introduces a novel hierarchical keyword taxonomy and an integrated assignment algorithm considering semantic similarity, bids, and conflicts for conference review management.

Findings

Semantic similarity improves reviewer-paper matching accuracy.

Automatic conflict detection enhances review integrity.

Balanced reviewer workload is maintained effectively.

Abstract

This paper proposes an architecture and assignment management model of a conference management system that performs a precise and accurate automatic assignment of reviewers to papers. The system relies on taxonomy of keywords to describe papers and reviewers' competences. The implied hierarchical structure of the taxonomy provides important additional information - the semantic relationships between the separate keywords. It allows similarity measures to take into account not only the number of exactly matching keywords between a paper and a reviewer, but in case of non-matching ones to calculate how semantically close they are. Reviewers are allowed to bid on the papers they would like to (or not like to) review and to explicitly state conflicts of interest (CoI) with papers. An automatic CoI detection is checking for additional conflicts based on institutional affiliation, co-authorship (within the local database) and previous co-authorship in the past (within the major bibliographic indexes and digital libraries). The algorithm for automatic assignment takes into account all - selected keywords, reviewers' bids and conflicts of interest and tries to find the most accurate assignment while maintaining load balancing among reviewers.