| paper | d4-conf-aaai-2014-0300 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | unclear |
The paper’s FUTUREMATCH findings are simulation results without a stated computational guarantee, and they are not numbered theorem-type results. Its NP-hardness claim is attributed to earlier work, not established as a numbered result of this paper, so no eligible computational result anchors a mirror.
fails bit a — no named computational result to mirror
The paper’s FUTUREMATCH framework, matching objectives, and simulation findings are empirical; the NP-hardness claim is attributed to prior work.
I can’t make a case that meets your anchor requirement: this paper states no numbered theorem, lemma, corollary, or proposition asserting a computational result. Its only complexity claim is that the standard clearing problem with finite cycle cap \(L>2\) is NP-hard, attributed to Abraham, Blum, and Sandholm (2007). That is an unnumbered claim about a result from elsewhere, not an eligible anchor proved or cited as a numbered result here.
So there is no per-anchor continuous question to state without inventing an anchor. The dynamic matching framework and its experimental findings may motivate a population-continuous model, but they do not supply the named computational result this case must rest on.
The proponent raised no result-level anchor, so there is nothing to defeat one by one. Under the stated standard, no continuous mirror has yet been established. The paper’s own contribution is FUTUREMATCH, a learning framework supported by simulation results; its only complexity claim is that finite cycle clearing is NP-hard, attributed to earlier work. It proves no numbered computational result whose continuous counterpart this case can assess.
The strongest objection to a proposed mirror is that kidney exchange is built around a particular compatibility graph: which individual pairs can exchange kidneys, which cycles and chains are feasible, and how failures affect each planned exchange. A type description rich enough to preserve those facts may give nearly every pair a distinct type, leaving no useful reduction in support size. A coarse description by blood type, sensitization, or other medical classes can produce a finite population model, but aggregate compatibility rates may not preserve feasible disjoint cycles, chain structure, or expected transplants. Recovering those details could require a new stochastic graph model, rather than a continuization of a result in this paper.
That objection does not establish the universal claim. A mass-allocation model over recurring patient-donor cohorts, with type-level compatibility and failure rates, is plausible; cycles, chains, and fairness could be reformulated for it. The paper’s national-exchange setting gives such a regime some credibility. Whether that model is tractable or faithful would need investigation, and neither difficulty alone makes the question unworthy. So the defensible negative is narrower: this paper offers a weak target for the programme’s result-by-result continuization exercise. I cannot honestly rule out a worthwhile continuous mirror in every scenario.
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.