| paper | d4-conf-ijcai-2021-0008 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | yes |
The decisive failure is bit (a): the efficient algorithm for socially optimal EF1 allocation appears only as a prose summary of Benabbou et al. (2020), without a numbered theorem, lemma, corollary, or proposition in this paper. The proposed type-mass allocation is a plausible extension of that topic, but it cannot supply the required named computational result from this paper.
fails bit a — no named computational result to mirror
The proposed variant concerns only the prose claim about welfare-maximizing EF1 allocation under binary submodular valuations; it leaves the paper's explanation, privacy, housing-diversity, group-fairness, and data-driven game-learning work aside.
There is no eligible anchor in this paper: it contains no numbered Theorem, Lemma, Corollary, or Proposition asserting a computational result. It does summarize computational findings in prose, including an efficient algorithm for socially optimal EF1 allocation under binary submodular valuations, attributed to Benabbou et al. (2020). But that summary has no result number printed here, so I cannot honestly present it as an anchor under your rules.
The strongest *non-qualifying* lead would be a continuous version of that allocation problem. Let \(T\) be a finite set of household valuation types, \(G\) a set of apartment categories, and \(\mu_t\) the fraction of households of type \(t\). Given per-capita supplies \(q_g\), choose masses \(x_{t,S}\) of type-\(t\) households receiving each integral bundle \(S\subseteq G\). Require \(\sum_S x_{t,S}=\mu_t\), respect each supply \(q_g\), and require EF1 between every pair of bundles used with positive mass; maximize total welfare. Large housing cohorts with repeated valuation types make this a plausible population-level question. The paper’s stated algorithmic claim suggests tractability as a conjecture, but it does not establish the continuous result.
So there is a plausible mirror to investigate, but no case that meets the required numbered-anchor standard.
The strongest negative case is formal: this paper supplies no numbered theorem, lemma, or corollary stating a computational result. The only candidate, the efficient socially optimal EF1 algorithm for binary submodular valuations, is a prose summary of Benabbou et al. (2020), not a numbered result in this paper. By the stated anchor standard, there is no eligible result to mirror.
The proposed mass-allocation model does not fix that gap. It is a coherent high-multiplicity variant, but it chooses replicated apartment categories, per-capita supplies, and an EF1 condition across every bundle pair in the allocation’s support. Those choices define a new support-sensitive allocation problem; they are not specified by the result summarized here. Restoring block-specific apartments and scaled ethnic quotas could make the housing scenario closer to the paper, but would still be a new continuous problem rather than a mirror of a computational theorem presented in it.
That is a case against this paper as an anchor, not proof that no worthwhile mirror exists in any scenario. The proposed model can plausibly be made into one, so the broader negative claim is weak; I cannot honestly say that every repaired version fails.
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.