| paper | d4-conf-ijcai-2015-0255 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | medium |
| authors would recognise it | yes |
P→P discrete result → its continuous mirror
Answer: Polynomial-time solvable in \(m,\tau,L\): sort items by decreasing expected positional weight \(s_i=\sum_{t=1}^{\tau}\mu_t w_{t(i)}\), using exact integer arithmetic. · discrete side: polynomial · adjudication: proved
full proof & adversarial review →
Proposition 3
statement extracted from the paper’s text layer
Given \(\tau\) ranking types \(t\in S_n\) with positive rational masses \(\mu_t\) summing to \(1\), and fixed strictly decreasing rational weights \(w\), find a ranking \(\rho\in S_n\) minimizing \(\sum_t \mu_t d_{PS}(\rho,t)\).
The society is a distribution \(\mu\) over complete ranking types. The decision variable is one centroid ranking \(\rho\); the objective is its expected positional Spearman distance, \(\sum_t\mu_t d_{PS}(\rho,t)\). Equivalently, sort items by their mass-weighted positional scores.
Rational masses can be cleared to a finite repeated-ranking profile, so the mirror may add no new optimizer behavior; this bounds its novelty but does not defeat the intended high-multiplicity question.
fatal: False
The mirror covers the weighted centroid computation for strictly decreasing positional scoring weights in Proposition 3, including the Borda–Spearman case in Proposition 1. It leaves the paper’s other distance characterizations, biased scoring rules, incomplete-ranking experiments, and joint \(k\)-cluster optimization outside this anchor.
The paper has no named computational result that meets the anchor requirement. Its numbered results are propositions about which distances characterize which aggregation rules, or about whether a distance can have certain metric properties. For example, Proposition 3, proved here, shows that positional Spearman distance with strictly decreasing weights characterizes the scoring rule with those weights. Proposition 1, attributed to Marden (1996), gives the Borda–Spearman connection. Neither asserts a complexity classification. The paper’s \(O(n\log n)\) running-time claim appears in the prose, not in a numbered result.
There is a plausible continuous-population angle: treat each complete ranking as a type, its rational mass as its share of a large recommender population, and ask for a centroid ranking minimizing the population’s expected positional Spearman distance. That is a natural high-multiplicity version of the paper’s centroid computation. But presenting it as an anchored positive case would stretch the instructions: the relevant proposition establishes a characterization, not a named computational result. So I cannot supply a qualifying anchor or a prover-ready question per anchor.
The strongest negative point is that the paper has no numbered result about computational complexity to mirror. Propositions 1–10 characterize aggregation rules by distances or rule out certain metric properties. The \(O(n\log n)\) running-time claim is in the prose, not a complexity result. So none of those propositions can serve as a qualifying anchor for a complexity mirror.
That point does not defeat the stronger clustering mirror. For a distribution \(\mu\) over ranking types \(T=S_n\), one can ask for \(k\) centroids \(\rho_1,\ldots,\rho_k\) minimizing
\[
\sum_{t\in T}\mu_t\min_{z\in\{1,\ldots,k\}} d(\rho_z,t).
\]
This is the population version of the paper’s clustering objective, and the paper explicitly presents that objective and an algorithm for it. It needs no named theorem to make it a computational question. Clearing the denominators of rational masses recovers a finite profile with repeated rankings; that is high multiplicity, which supports the mirror rather than counting against it. A type-level assignment can also split mass across clusters, as the continuum model permits.
So the case against is weak beyond the anchoring objection: the proponent did not supply a qualifying named-result anchor, but the paper still offers a plausible population-clustering problem to study. I cannot honestly argue that no worthwhile mirror exists in any scenario without leaning on objections the brief rules out.
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.