| paper | d4-conf-ecai-2016-0277 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
Theorem 1 has a natural analogue over a distribution of ranking types, but it does not assert a computational result. The experiments report empirical frontier statistics without a stated guarantee, so the paper supplies no qualifying result to anchor a computational mirror.
fails bit a — no named computational result to mirror
The proposed mirror translates Theorem 1's subgroup persistence into inclusion of ranking-type supports. It leaves aside Section 4's empirical frontier ratios, saturation estimates, and false-positive measurements.
There is a clean continuous counterpart to the paper’s central theorem, but there is no named computational result to anchor a ChoCo case to. The only numbered result is Theorem 1, proved in this paper: for strict, transitive, complete preferences, if \(A\subset A'\), then every outcome Pareto optimal for \(A\) is Pareto optimal for \(A'\), or \(O_A^*\subseteq O_{A'}^*\). That is a structural result about Pareto sets, not a complexity classification, algorithmic guarantee, or computational formulation. The experiments in Section 4 compute Pareto sets, but they are not numbered computational results. So Theorem 1 is the strongest structural hook, yet it does not meet the requested anchor criterion.
The natural population model is still quite direct. Let an outcome be a choice from a fixed, explicitly listed slate \(O\). A type \(t\) is a complete strict ranking of that slate. A continuous society is a distribution \(\mu\) over \(\tau\) such types, where \(\mu_t\) is the fraction of the population with ranking \(t\). This fits, for example, a large membership organization choosing among policy packages when members follow a small set of standardized constituency mandates: the number of members \(N\) is much larger than the number of distinct rankings \(\tau\). The relevant adjustment to Pareto improvement in a continuum is that an alternative must be weakly better for almost everyone and strictly better for a positive mass of members.
A precise candidate is Continuous Pareto-Frontier Enumeration. Given the explicit set \(O\) and a binary-rational distribution \(\mu\) over strict-ranking types, output the exact set
\[
P(\mu)=\{o\in O:\nexists o'\in O\text{ such that every type in }\operatorname{supp}(\mu)\text{ strictly prefers }o'\text{ to }o\}.
\]
The solution is the full frontier; there is no scalar objective to optimize. Theorem 1 has the exact support-based analogue: if \(\operatorname{supp}(\mu)\subseteq\operatorname{supp}(\nu)\), then \(P(\mu)\subseteq P(\nu)\). Adding positive mass of new ranking types can only rule out more potential dominators. Changing the masses while keeping the support fixed does not change the frontier.
For this explicitly represented problem, a direct pairwise scan computes the frontier in \(O(\tau |O|^2)\) ranking comparisons, so the expected outcome is tractability. The paper’s subgroup result is relevant: a frontier found on a subgroup’s types remains valid as further types join, under the strict-preference assumption. Its open-environment discussion makes that use recognizable to the authors, even though their experiments use small groups.
The weak point is that this gives a structural population analogue, not a paper-anchored complexity result. In this mirror, fractional masses have no effect beyond whether a type has positive mass, and the explicit-outcome computation is already straightforward. Making outcomes succinctly represented could create harder computational questions, but the paper does not study that representation. Thus the strongest honest positive claim is that Theorem 1 survives a natural high-multiplicity interpretation and suggests an exact frontier-enumeration problem; it does not supply a qualifying computational anchor for the programme.
I would reject this paper as a ChoCo anchor, though Theorem 1 does have a correct structural analogue. The paper contains no numbered result about the complexity of a computational problem: Theorem 1 is a set-inclusion claim, and the experiments report frontier statistics rather than an algorithmic guarantee or complexity result. The proposed continuous frontier therefore does not supply the missing computational result.
For the proponent’s finite ranking model, the analogue is
\[
P(\mu)=\{o\in O:\nexists o'\in O\ \forall t\in\operatorname{supp}(\mu),\ o'\succ_t o\}.
\]
It is valid, but depends only on which types have positive mass. Two societies with the same support have the same Pareto frontier, whether a type has mass \(10^{-100}\) or \(1/2\). That follows from Pareto’s unanimity condition: any positive-mass type can veto a dominator, regardless of how small its share is. With finitely many outcomes, even distributions over utility descriptions collapse to finitely many preference order types for this purpose. The continuous mass vector adds no computational content.
The stronger scenario—a large organization with many members sharing a small set of mandate rankings—makes high multiplicity sensible, but it does not rescue the mirror. For an explicitly listed outcome set, the frontier can be computed by checking pairs of outcomes against the \(\tau\) supported rankings, in \(O(\tau |O|^2)\) comparisons. A succinct combinatorial outcome space might make the search difficult, but then the difficulty comes from an outcome representation the paper neither specifies nor analyzes. Choosing a subgroup to maximize frontier coverage, or allowing a fraction of members to lose in an approximate Pareto test, could introduce mass-sensitive computational questions; those add a new selection problem or a new efficiency criterion.
The paper’s subgroup experiments also do not yield a stable continuum target. If a subgroup is a fixed positive fraction of a large society with finite support, it will eventually contain every positive-mass type with high probability, so its frontier converges to the full society’s frontier. A subgroup kept small enough to omit types remains a finite-group experiment. Deliberately selecting which types to include avoids that limit, but makes the task discrete support selection.
So the best proposed anchor survives as a structural extension, and the high-multiplicity setting is plausible. The case against is narrower: neither yields a meaningful computational mirror of this paper’s result. I would not claim that no independently formulated problem in this area could be worthwhile; the paper leaves those problems unspecified.
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.