| paper | d4-conf-ecai-2014-0070 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
Propositions 1–4 state logical properties, and Section 4's five-step method is a proposed procedure whose correctness the authors leave open; no named result gives a computational guarantee or hardness result. The survey-panel count adds an outer aggregation over individual inferences, so it cannot supply the missing computational anchor.
fails bit a — no named computational result to mirror
No result is covered: Propositions 1–4 are logical properties, and the inference procedure in Section 4 has no named computational guarantee.
I can’t make a qualifying anchored case for this paper. It has no named computational result asserting that a problem is in P, NP-hard, FPT, or another complexity class.
The numbered propositions are logical properties: Proposition 1 says that desiring \(\phi\) above level zero means not desiring \(\neg\phi\); Proposition 2 says an impossible proposition has desire level \(1\); Proposition 3 says positive desire for \(\phi\) makes \(\phi\) maximally potentially desirable; and Proposition 4 gives an inference rule combining two conditional desires. None is a complexity or algorithmic result. Section 4 does outline a five-step procedure for inferring desires, but the conclusion calls the approach preliminary and says its formal correctness still needs proof.
The most plausible population extension would be a large survey panel with many respondents but few distinct, complete conditional-desire profiles. A type would be a profile together with its interpretation of the relevant context; its mass would be the fraction of respondents with that type. Given a context \(\chi\), a query \(\psi\), and a threshold \(q\), one could ask whether at least mass \(q\) of the panel’s types would infer \(\psi\) above their \(\Delta\)-inconsistency level using the paper’s procedure. That is a reasonably precise high-multiplicity question, but it adds a population aggregation task to the paper’s single-agent inference problem. No named result here anchors it, and the paper gives no basis for predicting whether it is tractable or hard.
The paper’s graded values over worlds are not population mass. Treating those values as a continuous society would miss the programme’s central requirement. The survey-panel extension is therefore only a possible new research question, not an honest continuous mirror of a named computational result in this paper.
The strongest objection is that this paper gives no computational result to continuize. Propositions 1–4 state logical properties and inference rules for one agent’s desires. Section 4 sketches a five-step inference procedure, but the authors say its correctness still needs a formal proof; they give no complexity claim for it or for a named computational problem. That leaves no result here for a continuous mirror to sharpen or reclassify.
The survey-panel proposal is the best direct population extension. Let each type be a complete desire base and context, and let \(\mu_t\) be its share. For a query \(\psi\), one could calculate the mass of types whose inference derives \(\psi\). This is well-posed, and repeated desire profiles make a plausible high-multiplicity regime. But with explicit support, the calculation applies the paper’s inference procedure to each of the \(\tau\) types and sums their masses. The population layer adds no mass-dependent inference or optimization question. Studying the procedure’s complexity could be useful, but that would be a new analysis of an unproved procedure, not a mirror of a result in this paper.
A richer proposal might define collective desires, or ask how much population mass must be persuaded before a proposition crosses a group threshold. That could become a worthwhile computational social choice problem. The paper supplies neither a collective aggregation rule nor actions and costs for changing desire types. One can avoid comparing graded desire strengths by counting agents who infer \(\psi\), but then the campaigning problem still comes from added choices about interventions and costs. It is a new model built around the paper’s inference machinery, not a continuous version of one of its results.
The negative case is strongest on the absence of a computational anchor, not on the claim that no plausible population of agents could share desire types. I cannot rule out a worthwhile follow-on project. I would argue against commissioning a continuous mirror of this paper: its most natural population lift is an outer weighted count, while the more ambitious versions require the authors to supply a new collective problem first.
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.