| paper | d4-conf-ecai-2010-0109 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The paper has no numbered Theorem, Lemma, Corollary, or Proposition asserting a computational result. Proposition 1 states a fairness property, while the allocation formulations and solver experiments are not numbered results; the cited hardness claims are unnumbered and attributed to prior work. The proposed continuous Gini allocation is coherent, but no computational result of this paper anchors it.
fails bit a — no named computational result to mirror
The proposed mirror covers paper assignment under the generalized Gini welfare criterion; it leaves the maxmin, leximin, and Choquet models untreated. It is not anchored to a qualifying computational result in the paper.
The strongest honest case for a continuous mirror is about high-volume paper assignment: imagine a reviewing network with many papers in a small number of topic categories and many reviewers in a small number of expertise and workload types. Reviewers of a type have the same utility for each paper category and the same workload limits. Mass represents the fraction of reviewers in each type. This is a plausible high-multiplicity regime, and it mirrors the paper’s generalized-Gini allocation model.
But the paper has no qualifying named computational anchor under your rule. Its only numbered proposition, Proposition 1, is a fairness property: for a convex capacity \(v\), if \(x_1,\ldots,x_p\) all have the same Choquet integral, then every convex combination of them has Choquet integral at least that common value. It does not assert tractability or hardness. The NP-hardness claims about maxmin and leximin allocation appear in unnumbered prose and cite other work. The formulations \(\Pi'_W\) and \(\Pi''_{Cv}\) are computed and tested, but are not stated as numbered theorems or propositions.
If an unanchored candidate is useful, one precise mirror of the paper’s Gini model is this. Let \(T\) be reviewer types with rational masses \(\mu_t\) summing to \(1\), and \(P\) paper categories with per-capita supplies \(\rho_p\). Type \(t\) has additive utilities \(u_{tp}\). Let \(B_t\) be the integer paper bundles allowed by its workload limits. Choose \(y_{t,b}\ge 0\), the mass of type \(t\) receiving bundle \(b\), subject to
\[
\sum_{b\in B_t}y_{t,b}=\mu_t,\qquad
\sum_{t,b}y_{t,b}b_p=r_p\rho_p\quad(p\in P),
\]
where \(r_p\) is the required number of reviews per paper in category \(p\). Each bundle remains integral; the mass of a reviewer type can be split among bundles. The resulting population utility distribution puts mass \(y_{t,b}\) at \(\sum_p u_{tp}b_p\). The problem is to find a feasible \(y\) maximizing its continuous Gini welfare
\[
\int_0^1 2(1-q)Q_y(q)\,dq,
\]
where \(Q_y\) is the ascending utility quantile.
This candidate covers the paper-assignment/Gini angle, not its other welfare models. Its weakest point is that it is a proposed extension of an unnumbered formulation, not a mirror of a qualifying named result. The paper’s CPLEX experiments also do not establish a complexity classification for it.
This paper has no qualifying computational result to anchor a ChoCo mirror. Proposition 1 is a fairness property, not a complexity result. The Gini and Choquet programs are reformulations, and the tables report solver performance; neither establishes a complexity classification. The maxmin and leximin hardness claims appear in unnumbered prose and are attributed to earlier work. Under the stated standard, there is no paper result here for a mirror to preserve or re-examine.
The strongest candidate is still coherent: a large reviewing network with finitely many reviewer types and paper categories, where each reviewer receives an integral bundle, type mass can be split across bundles, and a continuous Gini objective evaluates the resulting utility distribution. That is a sensible high-multiplicity model. But it is a new computational problem built from the paper’s allocation formulation, not a mirror of a computational result the paper proves. Its value would have to be judged as a separate project.
That distinction makes the case against commissioning this *paper as a result-anchored mirror* strong. It does not support the broader claim that no worthwhile continuous question exists in any scenario. Reviewer types can recur, the Gini objective does not require named identities, and the proposed mass model does not degenerate. So the honest negative case is decisive on the missing anchor, but weak as a universal objection to the continuous model itself.
The adversarial triple: the proponent anchors on up to three named results; the opponent sees that case and must defeat every anchor; the judge decides which case convinced it. These are the pipeline’s own outputs, generated by tools/triple_run.py — no human edited them. The paper’s own text is not reproduced here beyond the quoted statement above.