Pith. sign in

REVIEW 2 major objections 5 minor 2 cited by

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

T0 review · 2 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper argues that group fairness under submission limits is an efficiently solvable LP, while individual fairness is NP-hard.

desk verdict One correct impossibility theorem buried under a false LP-equivalence claim and an invalid NP-hardness reduction; the paper's central mechanism does not work. read the letter →

arxiv 2502.00690 v1 pith:G7SXKV5M submitted 2025-02-02 cs.LG cs.AIcs.CYcs.DL

Yuefan Cao , Xiaoyu Li , Yingyu Liang , Zhizhou Sha , Zhenmei Shi , Zhao Song , Jiahao Zhang This is my paper · ORCID
classification cs.LGcs.AIcs.CYcs.DL MSC 68Q2590C0590C10
keywords deskrejectionsubmissionlimitsgroupfairnessindividuallinearprogrammingintegerNP-hardnessAIconferences
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper analyzes what happens when AI conferences cap the number of submissions per author and then desk-reject papers that push authors over the cap. It first shows that an 'ideal' rejection scheme, one that removes exactly each author's excess papers and touches no one else, cannot exist once there are three or more authors. It then defines two fairness objectives: individual fairness, the largest fraction of any author's papers lost, and group fairness, the average fraction lost across authors. The central results are that minimizing individual fairness is NP-hard, while minimizing group fairness is claimed to be solvable exactly by one linear program, because the LP relaxation is said to have a 0-1 optimal solution. If that equivalence is right, a conference could desk-reject papers in a way that provably minimizes the average harm to authors, instead of relying on submission order.

What carries the argument

The load-bearing object is the author-paper incidence matrix $W \in \{0,1\}^{n \times m}$, where $W_{i,j}=1$ if author $i$ is a coauthor of paper $j$, together with the diagonal normalization $D=\operatorname{Diag}(|P_1|,\dots,|P_n|)$. The group-fair objective is to maximize $\mathbf{1}_n^\top D^{-1} W r$ subject to $W r \le x \mathbf{1}_n$ and $r \in [0,1]^m$, where $r_j$ is the fraction of paper $j$ that is kept. The proof of Theorem B.9 is the step that carries the positive result: it invokes the fundamental theorem of linear programming to argue that an optimal extreme point of this polytope has coordinates in $\{0,1\}$, so the relaxation and the integer program have the same optimum. The cost function $c(a_i,S)=(|P_i|-(Wr)_i)/|P_i|$ converts those counts into the fairness losses the LP minimizes.

What would settle it

Solve the triangle instance with three authors, three papers, and $x=1$; the point $r=(1/2,1/2,1/2)$ is feasible, is an extreme point of the relaxed polytope, and has objective $3/2$, while every integral feasible vector has objective at most $1$, so the claimed LP/integer equivalence fails on this instance.

Watch

Extended reading notes

Core claim

The central discovery is a computational dichotomy inside a single administrative problem. The paper defines a cost for each author as the fraction of their submitted papers that are desk-rejected when a set of papers is kept, and it contrasts the maximum cost (individual fairness) with the average cost (group fairness). It proves the ideal desk-rejection target is unattainable for $n \ge 3$ authors, proves individual fairness minimization is NP-hard via a reduction from Set Cover, and proves (Theorem 5.16/B.9 of the paper) that the group fairness problem can be relaxed to a linear program whose optimal solution is still a 0-1 vector, so an off-the-shelf LP solver finds an exactly optimal desk-rejection set. On its own example, this LP-based selection is strictly fairer than the order-based rule that several conferences use.

Load-bearing premise

The argument depends on the claim that every extreme point of the relaxed feasible polytope $\{r \in [0,1]^m : W r \le x \mathbf{1}_n\}$ has all coordinates 0 or 1, so solving the LP gives a genuine set of papers to reject.

Editorial extensions

If this is right

  • Under the paper's equivalence, a conference adopting this LP mechanism would desk-reject in polynomial time while matching the integer program's optimum.
  • Because the group-fair objective is a lower bound on the individual-fairness objective, an LP-optimal desk-rejection set is also a necessary first step toward controlling the worst-case harm suffered by any single author.
  • The paper's case study shows that the order-based rejection method used by several top conferences is strictly less fair than the LP solution, since it can reject the only paper of an author who coauthored the 26th submission of a prolific colleague.
  • The NP-hardness result implies that no efficient exact algorithm can guarantee the worst-case fairness metric unless $\mathrm{P} = \mathrm{NP}$.
  • At the scale of major conferences, with on the order of $10^4$ submissions, the LP formulation can be solved with standard solvers, making the mechanism practically deployable.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's own three-author, $x=1$ negative example doubles as a stress test for the LP claim: the fractional point $r=(1/2,1/2,1/2)$ is an extreme point of the relaxed polytope and beats every integral solution, so the claimed equivalence would need a different argument or a rounding rule.
  • A practical fallback is to solve the LP and then round the solution, but the optimality guarantee would have to be re-proved, and the rounding bias should be audited on author counts.
  • The same ideal-rejection impossibility applies to any quota system with overlapping memberships, so the group-fair LP formulation is a template for enforcing caps on grants, courses, or reviewer workloads without collateral damage.
  • Because the cost function averages harm over authors, group fairness implicitly treats papers with fewer coauthors as cheaper to reject and may systematically target small collaborations; distributional audits should accompany deployment.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

2 major / 5 minor

Summary. This paper studies fairness of desk-rejection policies under per-author submission limits. It formalizes an ideal desk-rejection criterion, proves impossibility for n≥3 authors, introduces individual and group fairness metrics, claims individual fairness optimization is NP-hard and group fairness optimization reduces to an LP with an optimal 0-1 solution, and presents case studies including a comparison with the CVPR 2025 mechanism. The main theoretical pivot is Theorem 5.16/B.9, which asserts equivalence between the LP relaxation in Definition 5.15 and the integer program in Definition 5.13.

Significance. The problem framing and the impossibility example for the ideal desk-rejection criterion are useful observations, and the topic is timely. However, the central LP-equivalence claim is false, and the NP-hardness proof is flawed; the advertised contribution of efficiently optimizing group fairness with an off-the-shelf LP solver is not established. If the LP claim could be repaired, the group-fairness approach would be of practical interest, but as it stands the main theoretical results do not hold.

major comments (2)
  1. [Theorem 5.16 / Appendix B.3 (Definition 5.15)] The central equivalence claim is false. On the instance from Lemma A.6 with n=3, m=3, x=1 and papers p1=(a1,a2), p2=(a1,a3), p3=(a2,a3), the LP in Definition 5.15 is maximize r1+r2+r3 subject to r1+r2≤1, r1+r3≤1, r2+r3≤1, and r∈[0,1]^3. The point r=(1/2,1/2,1/2) is an extreme point of the feasible polytope, and the LP optimum is 3/2, whereas the integer program in Definition 5.13 has optimum 1. Thus the relaxed and integral optima differ, contradicting Theorem 5.16/B.9. Consequently Algorithm 1's transformation step (lines 19–24) is undefined for fractional LP solutions, and the advertised efficient group-fairness optimization is not valid.
  2. [Theorem B.7 in Appendix B.2] The NP-hardness reduction adds a cardinality constraint ∥r∥1 ≤ K that is not present in the Individual Fairness-Aware Submission Limit Problem in Definition 5.7 or its matrix form in Definition 5.9. Reducing Set Cover to this augmented problem does not establish hardness of the original problem. The sentence 'this problem is easier than the optimization problem defined in Definition 5.7' is also incorrect: adding a constraint makes the feasible set smaller, and the reduction must produce an instance of the original problem. Therefore the proof does not establish Theorem 5.11.
minor comments (5)
  1. [Proposition 5.6 / Proposition B.1] The inequality in Proposition 5.6 is stated as ζind(S) ≤ ζgroup(S), but the proof in Appendix B.1 establishes ζgroup(S) ≤ ζind(S); the direction in the main text should be corrected.
  2. [Algorithm 1, line 8] The condition 'if pj ∈ Ai' is inconsistent with the notation in Definition 3.1 and Definition 5.8; it should be 'if ai ∈ Aj'.
  3. [Algorithm 1, lines 19–24] The transformation step only selects papers with rj = 1. Since LP solvers may return fractional extreme points or interior solutions, an integrality guarantee or an explicit rounding procedure is required.
  4. [Lemma A.7] The proof examines only the sequential rejection of p1 then p2 and concludes failure, but it does not rule out other rejection sets. A rigorous argument is needed to show that no subset of rejected papers can give each author exactly n−2 remaining papers.
  5. [Lemma A.4 and Lemma A.5] Both lemmas state 'Let n = 1' in their conditions although they are about the n=2 case.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the central derivation is self-contained, and the main theorem's failure is a mathematical error, not a circular argument.

full rationale

The paper's derivation chain does not reduce to its own assumptions. The submission-limit problem, the ideal desk-rejection definition, and the two fairness metrics are defined independently, and the positivity/negativity results in Section 4 are proved directly from those definitions. The group-fairness LP claim in Theorem 5.16/B.9 is argued from the fundamental theorem of linear programming rather than from the conclusion itself; the fact that the integrality assertion is false on the paper's own triangle instance is a correctness failure, not circularity. The only visible self-citations appear in Remark 5.17, where the authors cite their own LP solver papers for runtime complexity, but this is tangential to the core equivalence claim and does not carry the main argument. No fitted parameter is relabeled as a prediction, no uniqueness theorem is imported from the authors' prior work, and no known result is merely renamed. Accordingly, the paper receives a low circularity score despite its substantive mathematical flaws.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

No fitted parameters or invented physical entities appear. The model's inputs are the author-paper incidence matrix, the submission limit x, and the defined fairness costs; x is part of the problem statement, not a fitted constant. The load-bearing unstated assumption is the integrality of LP extreme points, which fails on the paper's own triangle construction.

assumptions (3)
  • ad hoc to paper The feasible polytope {r in [0,1]^m : Wr <= x 1_n} has 0-1 extreme points.
    Invoked in the proof of Theorem B.9 to conclude the LP optimum is an integer solution. This is false for general 0-1 matrices W; the triangle instance of Lemma A.6 gives a fractional extreme point at r=(1/2,1/2,1/2).
  • standard math Set Cover is NP-hard.
    Used in Theorem B.7 as the source problem for reduction. The cited hardness is correct, but the reduction itself is invalid because it adds a cardinality constraint not present in the target problem.
  • standard math The fundamental theorem of linear programming: an LP optimum occurs at an extreme point.
    Correct as stated, but it does not imply that every coordinate of an extreme point is 0 or 1. The proof of Theorem B.9 over-applies this theorem.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dissecting Submission Limit in Desk-Rejections: A Mathematical Analysis of Fairness in AI Conference Policies." pith.science (2026). https://pith.science/paper/G7SXKV5M

@misc{pith2026250200690,
  author       = {Pith},
  title        = {Pith review of: Dissecting Submission Limit in Desk-Rejections: A Mathematical Analysis of Fairness in AI Conference Policies},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/G7SXKV5M}},
  note         = {Machine review of arXiv:2502.00690}
}
read the original 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.

Figures

Figures reproduced from arXiv: 2502.00690 by the authors.

Figure 1
Figure 1. The Matthew Effect in the AI community. This figure illustrates the worsening Matthew [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. The unfairness of desk rejection based on submission limits. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Our research objective. This figure presents the goal of our study: creating a more [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 2 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Force Matching with Relativistic Constraints: A Physics-Inspired Approach to Stable and Efficient Generative Modeling

    cs.LG 2025-02 reject novelty 4.0 of 10

    Force Matching replaces velocity matching in flow-based generative models with a relativistic force objective, but the toy experiments are designed so the model class matches the data generator exactly.

  2. Universal Approximation of Visual Autoregressive Transformers

    cs.LG 2025-02 reject novelty 4.0 of 10

    The paper's headline claim that VAR transformers universally approximate all Lipschitz image maps is not supported, because the theorem restricts the target class and its key lemma has an invalid linearity step.

Reference graph

Works this paper leans on

54 extracted references · 41 canonical work pages · cited by 2 Pith papers

  1. [1]

    Gpt-4 technical report

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 , 2023

    arXiv 2023

  2. [2]

    Scientific labor market and firm-level appropriation strategy in artificial intelligence research

    Nur Ahmed. Scientific labor market and firm-level appropriation strategy in artificial intelligence research. Technical report, MIT Sloan Working Paper, 2022

    work page 2022

  3. [3]

    Group fairness in peer review

    Haris Aziz, Evi Micha, and Nisarg Shah. Group fairness in peer review. Advances in Neural Information Processing Systems , 36, 2023

    work page 2023

  4. [4]

    Group fairness in peer review

    Haris Aziz, Evi Micha, and Nisarg Shah. Group fairness in peer review. Advances in Neural Information Processing Systems , 36, 2024

    work page 2024

  5. [5]

    The claude 3 model family: Opus, sonnet, haiku, 2024

    Anthropic. The claude 3 model family: Opus, sonnet, haiku, 2024

    2024

  6. [6]

    Desk rejecting: A better use of your time

    Ben W Ansell and David J Samuels. Desk rejecting: A better use of your time. PS: Political Science & Politics , 54(4):686--689, 2021

    work page 2021

  7. [7]

    The compute divide in machine learning: A threat to academic contribution and scrutiny? arXiv preprint arXiv:2401.02452 , 2024

    Tamay Besiroglu, Sage Andrus Bergerson, Amelia Michael, Lennart Heim, Xueyun Luo, and Neil Thompson. The compute divide in machine learning: A threat to academic contribution and scrutiny? arXiv preprint arXiv:2401.02452 , 2024

    arXiv 2024

  8. [8]

    Stable video diffusion: Scaling latent video diffusion models to large datasets

    Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram Voleti, Adam Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets. arXiv preprint arXiv:2311.15127 , 2023

    arXiv 2023

Show all 54 references
  1. [9]

    The matthew effect in science funding

    Thijs Bol, Mathijs De Vaan, and Arnout van de Rijt. The matthew effect in science funding. Proceedings of the National Academy of Sciences , 115(19):4887--4890, 2018

    2018

  2. [10]

    Time to rethink the publication process in machine learning

    Yoshua Bengio. Time to rethink the publication process in machine learning. https://yoshuabengio.org/2020/02/26/time-to-rethink-the-publication-proce\ -in-machine-learning/, 2020

    2020

  3. [11]

    The simplex method of linear programming using lu decomposition

    Richard H Bartels and Gene H Golub. The simplex method of linear programming using lu decomposition. Communications of the ACM , 12(5):266--268, 1969

    1969

  4. [12]

    Solving linear programs in the current matrix multiplication time

    Michael B Cohen, Yin Tat Lee, and Zhao Song. Solving linear programs in the current matrix multiplication time. In Proceedings of the 51st Annual ACM SIGACT Symposium on Theory of Computing , pages 938--942, 2019

    2019

  5. [13]

    Solving linear programs in the current matrix multiplication time

    Michael B Cohen, Yin Tat Lee, and Zhao Song. Solving linear programs in the current matrix multiplication time. Journal of the ACM (JACM) , 68(1):1--39, 2021

    2021

  6. [14]

    Fairness in information access systems

    Michael D Ekstrand, Anubrata Das, Robin Burke, Fernando Diaz, et al. Fairness in information access systems. Foundations and Trends in Information Retrieval , 16(1-2):1--177, 2022

    2022

  7. [15]

    Ai-generated research paper fabrication and plagiarism in the scientific community

    Faisal R Elali and Leena N Rachid. Ai-generated research paper fabrication and plagiarism in the scientific community. Patterns , 4(3), 2023

    2023

  8. [16]

    Fairness

    Nissim Francez. Fairness . Springer Science & Business Media, 2012

    2012

  9. [17]

    The perceived fairness of selection systems: An organizational justice perspective

    Stephen W Gilliland. The perceived fairness of selection systems: An organizational justice perspective. Academy of management review , 18(4):694--734, 1993

    1993

  10. [18]

    Computers and Intractability: A Guide to the Theory of NP-Completeness , volume 174

    Michael R Garey and David S Johnson. Computers and Intractability: A Guide to the Theory of NP-Completeness , volume 174. freeman San Francisco, 1979

    1979

  11. [19]

    Towards long-term fairness in recommendation

    Yingqiang Ge, Shuchang Liu, Ruoyuan Gao, Yikun Xian, Yunqi Li, Xiangyu Zhao, Changhua Pei, Fei Sun, Junfeng Ge, Wenwu Ou, et al. Towards long-term fairness in recommendation. In Proceedings of the 14th ACM international conference on web search and data mining , pages 445--453, 2021

    2021

  12. [20]

    Bias and fairness in large language models: A survey

    Isabel O Gallegos, Ryan A Rossi, Joe Barrow, Md Mehrab Tanjim, Sungchul Kim, Franck Dernoncourt, Tong Yu, Ruiyi Zhang, and Nesreen K Ahmed. Bias and fairness in large language models: A survey. Computational Linguistics , pages 1--79, 2024

    2024

  13. [21]

    Auto-debias: Debiasing masked language models with automated biased prompts

    Yue Guo, Yi Yang, and Ahmed Abbasi. Auto-debias: Debiasing masked language models with automated biased prompts. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) , pages 1012--1023, 2022

    2022

  14. [22]

    Denoising diffusion probabilistic models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. Advances in neural information processing systems , 33:6840--6851, 2020

    2020

  15. [23]

    Video diffusion models

    Jonathan Ho, Tim Salimans, Alexey Gritsenko, William Chan, Mohammad Norouzi, and David J Fleet. Video diffusion models. Advances in Neural Information Processing Systems , 35:8633--8646, 2022

    2022

  16. [24]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 770--778, 2016

    2016

  17. [25]

    Effects of editorial peer review: a systematic review

    Tom Jefferson, Philip Alderson, Elizabeth Wager, and Frank Davidoff. Effects of editorial peer review: a systematic review. Jama , 287(21):2784--2786, 2002

    2002

  18. [26]

    A faster algorithm for solving general lps

    Shunhua Jiang, Zhao Song, Omri Weinstein, and Hengjie Zhang. A faster algorithm for solving general lps. In Proceedings of the 53rd Annual ACM SIGACT Symposium on Theory of Computing , pages 823--832, 2021

    2021

  19. [27]

    Reducibility among combinatorial problems

    Richard M Karp. Reducibility among combinatorial problems. Complexity of Computer Computations , pages 85--103, 1972

    1972

  20. [28]

    A conversation on artificial intelligence, chatbots, and plagiarism in higher education

    Michael R King and ChatGPT. A conversation on artificial intelligence, chatbots, and plagiarism in higher education. Cellular and molecular bioengineering , 16(1):1--2, 2023

    2023

  21. [29]

    Matching papers and reviewers at large conferences

    Kevin Leyton-Brown, Yatin Nandwani, Hedayat Zarkoob, Chris Cameron, Neil Newman, Dinesh Raghu, et al. Matching papers and reviewers at large conferences. Artificial Intelligence , 331:104119, 2024

    2024

  22. [30]

    User-oriented fairness in recommendation

    Yunqi Li, Hanxiong Chen, Zuohui Fu, Yingqiang Ge, and Yongfeng Zhang. User-oriented fairness in recommendation. In Proceedings of the web conference 2021 , pages 624--632, 2021

    2021

  23. [31]

    Towards personalized fairness based on causal notion

    Yunqi Li, Hanxiong Chen, Shuyuan Xu, Yingqiang Ge, and Yongfeng Zhang. Towards personalized fairness based on causal notion. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval , pages 1054--1063, 2021

    2021

  24. [32]

    Duplicate submission and dual publication: what is so wrong with them?, 2013

    Seth S Leopold. Duplicate submission and dual publication: what is so wrong with them?, 2013

    2013

  25. [33]

    Path finding methods for linear programming: Solving linear programs in o (vrank) iterations and faster algorithms for maximum flow

    Yin Tat Lee and Aaron Sidford. Path finding methods for linear programming: Solving linear programs in o (vrank) iterations and faster algorithms for maximum flow. In 2014 IEEE 55th Annual Symposium on Foundations of Computer Science , pages 424--433. IEEE, 2014

    2014

  26. [34]

    Linear and nonlinear programming , volume 2

    David G Luenberger, Yinyu Ye, et al. Linear and nonlinear programming , volume 2. Springer, 1984

    1984

  27. [35]

    A survey on bias and fairness in machine learning

    Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan. A survey on bias and fairness in machine learning. ACM computing surveys (CSUR) , 54(6):1--35, 2021

    2021

  28. [36]

    Investigating gender fairness of recommendation algorithms in the music domain

    Alessandro B Melchiorre, Navid Rekabsaz, Emilia Parada-Cabaleiro, Stefan Brandl, Oleg Lesota, and Markus Schedl. Investigating gender fairness of recommendation algorithms in the music domain. Information Processing & Management , 58(5):102666, 2021

    2021

  29. [37]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervision. In International conference on machine learning , p...

    2021

  30. [38]

    Ranking mobility and impact inequality in early academic careers

    Ye Sun, Fabio Caccioli, and Giacomo Livan. Ranking mobility and impact inequality in early academic careers. Proceedings of the National Academy of Sciences , 120(34):e2305196120, 2023

    2023

  31. [39]

    Denoising diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denoising diffusion implicit models. arXiv preprint arXiv:2010.02502 , 2020

    2010 arXiv

  32. [40]

    Research and development - ai index, 2024

    Stanford. Research and development - ai index, 2024

    2024

  33. [41]

    Plagiarism, duplicate publication, and duplicate submission: They are all wrong! IEEE Antennas and Propagation Magazine , 45(4):47--49, 2003

    WR Stone. Plagiarism, duplicate publication, and duplicate submission: They are all wrong! IEEE Antennas and Propagation Magazine , 45(4):47--49, 2003

    2003

  34. [42]

    Oblivious sketching-based central path method for linear programming

    Zhao Song and Zheng Yu. Oblivious sketching-based central path method for linear programming. In Proceedings of the 38th International Conference on Machine Learning , volume 139 of Proceedings of Machine Learning Research , pages 9835--9847, 2021

    2021

  35. [43]

    Chatgpt: Optimizing language models for dialogue

    John Schulman, Barret Zoph, Christina Kim, Jacob Hilton, Jacob Menick, Jiayi Weng, Juan Felipe Ceron Uribe, Liam Fedus, Luke Metz, Michael Pokorny, et al. Chatgpt: Optimizing language models for dialogue. OpenAI blog , 2(4), 2022

    2022

  36. [44]

    Establishing sensible and practical guidelines for desk rejections

    Jaime A Teixeira da Silva, Aceil Al-Khatib, Vedran Katavi \'c , and Helmar Bornemann-Cimenti. Establishing sensible and practical guidelines for desk rejections. Science and Engineering Ethics , 24:1347--1365, 2018

    2018

  37. [45]

    The state of the art in peer review

    Jonathan P Tennant. The state of the art in peer review. FEMS Microbiology letters , 365(19):fny204, 2018

    2018

  38. [46]

    Solving tall dense linear programs in nearly linear time

    Jan van den Brand, Yin Tat Lee, Aaron Sidford, and Zhao Song. Solving tall dense linear programs in nearly linear time. In Proceedings of the 52nd Annual ACM SIGACT Symposium on Theory of Computing , pages 775--788, 2020

    2020

  39. [47]

    Attention is all you need

    Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, ukasz Kaiser, and Illia Polosukhin. Attention is all you need. Advances in neural information processing systems , 30, 2017

    2017

  40. [48]

    Dolfin: Diffusion layout transformers without autoencoder

    Yilin Wang, Zeyuan Chen, Liangjun Zhong, Zheng Ding, Zhizhou Sha, and Zhuowen Tu. Dolfin: Diffusion layout transformers without autoencoder. arXiv preprint arXiv:2310.16305 , 2023

    2023 arXiv

  41. [49]

    Factors influencing perceived fairness in algorithmic decision-making: Algorithm outcomes, development procedures, and individual differences

    Ruotong Wang, F Maxwell Harper, and Haiyi Zhu. Factors influencing perceived fairness in algorithmic decision-making: Algorithm outcomes, development procedures, and individual differences. In Proceedings of the 2020 CHI conference on human factors in computing systems , pages...

    2020

  42. [50]

    Early-career setback and future career impact

    Yang Wang, Benjamin F Jones, and Dashun Wang. Early-career setback and future career impact. Nature communications , 10(1):4331, 2019

    2019

  43. [51]

    Joint multisided exposure fairness for recommendation

    Haolun Wu, Bhaskar Mitra, Chen Ma, Fernando Diaz, and Xue Liu. Joint multisided exposure fairness for recommendation. In Proceedings of the 45th International ACM SIGIR Conference on research and development in information retrieval , pages 703--714, 2022

    2022

  44. [52]

    Tokencompose: Text-to-image diffusion with token-level supervision

    Zirui Wang, Zhizhou Sha, Zheng Ding, Yilin Wang, and Zhuowen Tu. Tokencompose: Text-to-image diffusion with token-level supervision. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 8553--8564, 2024

    2024

  45. [53]

    Omnicontrolnet: Dual-stage integration for conditional image generation

    Yilin Wang, Haiyang Xu, Xiang Zhang, Zeyuan Chen, Zhizhou Sha, Zirui Wang, and Zhuowen Tu. Omnicontrolnet: Dual-stage integration for conditional image generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition , pages 7436--7448, 2024

    2024

  46. [54]

    Causal intervention for leveraging popularity bias in recommendation

    Yang Zhang, Fuli Feng, Xiangnan He, Tianxin Wei, Chonggang Song, Guohui Ling, and Yongdong Zhang. Causal intervention for leveraging popularity bias in recommendation. In Proceedings of the 44th international ACM SIGIR conference on research and development in information retr...

    2021

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.