| paper | d4-conf-ijcai-2020-0769 |
| authors | — |
| venue | ? |
| filed under | unclassified |
| judged by | gpt-6-luna / xhigh (triple__luna__xhigh__c6r1) |
| judge confidence | high |
| authors would recognise it | no |
The paper reports simulation comparisons of trust methods, but contains no numbered theorem, lemma, corollary, or proposition asserting a computational result. The proponent acknowledges that its marketplace proposal is not anchored in such a result, so no continuous mirror passes the stated gate.
fails bit a — no named computational result to mirror
There is no named computational result to anchor; the paper's findings are empirical comparisons of methods by (\mathrm{RFU}) in simulated settings.
I can’t give a compliant anchor-based positive case: the paper contains no numbered Theorem, Lemma, Corollary, or Proposition asserting a computational result. Its findings are empirical comparisons of trust methods using RFU; Table 4 and the ExpFun method are not complexity-result anchors. So there are no anchors here to restate or to pair with the required prover questions.
The most plausible population scenario would be a large online marketplace with many advisors grouped by their complete recommendation behavior, including how that behavior changes over time. A type’s mass would be its share of the advisor pool, and a trust system would use those recommendations and the truster’s direct history to choose a trustee, aiming to reduce expected RFU. That seems like a sensible high-multiplicity extension of the paper’s setting, but it is not a mirror anchored in a named result of this paper.
The proponent gives no anchor to defeat: the paper has no numbered theorem, lemma, or corollary asserting a computational result. Its headline findings—that indirect trust often makes direct trust redundant and that ExpFun performs best—come from RFU measurements over simulations. They are empirical comparisons, not complexity claims.
The strongest repair is a large marketplace with many advisors, represented by a distribution over advisor types, and a question about which trust combiner minimizes expected RFU. That is a plausible high-multiplicity regime; the paper’s ten-advisor setup is only an experimental choice, and it already varies the dishonest-advisor share. But this repair changes the paper’s task. A faithful type must preserve the advisor’s recommendation history, trust-network position, and changing behavior, since the methods use those details. Grouping advisors only by labels such as “honest” or “camouflaging” discards relevant information; tracking their individual states makes the model large without giving the paper a new computational result. If the question is instead to optimize a combiner over advisor distributions, that is a new robust-policy problem, not a mirror of a result this paper proves.
So the negative case is strong under ChoCo’s stated gate: this paper offers no computational complexity result to continuize, and the best population version extends its experimental study rather than exposing a complexity landscape. I cannot honestly claim that no continuous advisor-population model could be worthwhile in trust research. That is a plausible adjacent direction; it simply is not an anchor supplied by this paper.
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.