Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Dissecting Submission Limit in Desk-Rejections: A Mathematical Analysis of Fairness in AI Conference Policies

Yuefan Cao, Xiaoyu Li, Yingyu Liang, Zhizhou Sha, Zhenmei Shi, Zhao Song, Jiahao Zhang

TL;DR

This work analyzes fairness in desk-rejection policies under per-author submission caps in AI conferences. It proves that an ideal policy rejecting solely based on an author's excess submissions is impossible when the author set size satisfies $n\ge 3$, and introduces two fairness metrics based on an author-centric cost. The authors show that optimizing individual fairness is NP-hard, while group fairness can be efficiently solved via linear programming, yielding a fairness-aware desk-rejection mechanism. Case studies indicate the LP-based approach achieves greater equity than existing practices (e.g., CVPR 2025), highlighting practical routes to more just conference submission management. The results have implications for policy design and fair evaluation in rapidly growing AI research communities.

Abstract

As AI research surges in both impact and volume, conferences have imposed submission limits to maintain paper quality and alleviate organizational pressure. In this work, we examine the fairness of desk-rejection systems under submission limits and reveal that existing practices can result in substantial inequities. Specifically, we formally define the paper submission limit problem and identify a critical dilemma: when the number of authors exceeds three, it becomes impossible to reject papers solely based on excessive submissions without negatively impacting innocent authors. Thus, this issue may unfairly affect early-career researchers, as their submissions may be penalized due to co-authors with significantly higher submission counts, while senior researchers with numerous papers face minimal consequences. To address this, we propose an optimization-based fairness-aware desk-rejection mechanism and formally define two fairness metrics: individual fairness and group fairness. We prove that optimizing individual fairness is NP-hard, whereas group fairness can be efficiently optimized via linear programming. Through case studies, we demonstrate that our proposed system ensures greater equity than existing methods, including those used in CVPR 2025, offering a more socially just approach to managing excessive submissions in AI conferences.

Dissecting Submission Limit in Desk-Rejections: A Mathematical Analysis of Fairness in AI Conference Policies

TL;DR

This work analyzes fairness in desk-rejection policies under per-author submission caps in AI conferences. It proves that an ideal policy rejecting solely based on an author's excess submissions is impossible when the author set size satisfies , and introduces two fairness metrics based on an author-centric cost. The authors show that optimizing individual fairness is NP-hard, while group fairness can be efficiently solved via linear programming, yielding a fairness-aware desk-rejection mechanism. Case studies indicate the LP-based approach achieves greater equity than existing practices (e.g., CVPR 2025), highlighting practical routes to more just conference submission management. The results have implications for policy design and fair evaluation in rapidly growing AI research communities.

Abstract

As AI research surges in both impact and volume, conferences have imposed submission limits to maintain paper quality and alleviate organizational pressure. In this work, we examine the fairness of desk-rejection systems under submission limits and reveal that existing practices can result in substantial inequities. Specifically, we formally define the paper submission limit problem and identify a critical dilemma: when the number of authors exceeds three, it becomes impossible to reject papers solely based on excessive submissions without negatively impacting innocent authors. Thus, this issue may unfairly affect early-career researchers, as their submissions may be penalized due to co-authors with significantly higher submission counts, while senior researchers with numerous papers face minimal consequences. To address this, we propose an optimization-based fairness-aware desk-rejection mechanism and formally define two fairness metrics: individual fairness and group fairness. We prove that optimizing individual fairness is NP-hard, whereas group fairness can be efficiently optimized via linear programming. Through case studies, we demonstrate that our proposed system ensures greater equity than existing methods, including those used in CVPR 2025, offering a more socially just approach to managing excessive submissions in AI conferences.

Paper Structure

This paper contains 42 sections, 18 theorems, 33 equations, 3 figures, 3 tables, 2 algorithms.

Table of Contents

  1. Introduction
  2. Roadmap.
  3. Related Works
  4. Desk Reject Mechanism
  5. The Competitive Race in AI Publication
  6. Fairness System Design
  7. Preliminary
  8. Notations
  9. Problem Formulation
  10. The Desk Rejection Dilemma
  11. Ideal Desk-Rejection
  12. Hardness of Ideal Desk-Rejection
  13. Fairness-Aware Desk-Rejection
  14. Fairness Metrics
  15. Hardness of Individual Fairness-Aware Submission Limit Problem
  16. ...and 27 more sections

Key Result

Theorem 4.3

Let $n = |\mathcal{A}|$ denote the number of authors in Definition def:submit_limit_problem. We can show that

Figures (3)

  • Figure 1: The Matthew Effect in the AI community. This figure illustrates the worsening Matthew Effect in the AI community, where senior researchers tend to have a significantly higher number of submissions, while junior researchers have relatively few.
  • Figure 2: The unfairness of desk rejection based on submission limits. Left: A careless mistake. In this scenario, a young student submits the only paper, co-authored with a professor who submits numerous papers, and carelessly exceeds the submission limit. The paper, which may aim to apply to graduate programs, secure employment, or form a chapter of the thesis, is very important for the student but may not be for the professor. Right: The desk rejection. If the paper is desk-rejected due to submission limits, it poses a minor inconvenience to the professor, and the professor can shrug about it due to his remaining papers. However, it could have severe consequences for the students, as the paper is crucial for the student's future plans.
  • Figure 3: Our research objective. This figure presents the goal of our study: creating a more equitable desk-rejection system. Consider Professor A, who has carelessly submitted numerous papers exceeding the submission limit, collaborating with another senior researcher (Professor B) with many submissions, and a young student with only one paper. Our proposed system prioritizes desk-rejecting papers from authors with a large number of submissions first, thereby increasing the student’s chances of having their paper accepted. This approach aims to mitigate the disparity in the impact of desk rejections and promote fairness.

Theorems & Definitions (53)

  • Definition 3.1: Submission Limit Problem
  • Definition 4.1: Ideal desk-rejection
  • Remark 4.2
  • Theorem 4.3: Hardness of Ideal Desk-Rejection
  • proof
  • Definition 5.1: Cost Function
  • Remark 5.2
  • Example 5.3
  • Definition 5.4: Individual Fairness
  • Definition 5.5: Group Fairness
  • ...and 43 more