| paper | d4-conf-aaai-2017-0873 |
| 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 has no numbered theorem, lemma, corollary, or proposition asserting a computational result, so criterion (a) fails. Its EM procedure and experimental findings do not provide the required named anchor for a continuous mirror.
fails bit a — no named computational result to mirror
No named computational result is covered; the EM procedure and experiments remain unanchored.
The paper has no numbered Theorem, Lemma, Corollary, or Proposition asserting a computational result. It presents an EM inference algorithm and reports simulation and Mechanical Turk results, but neither provides a qualifying anchor for the continuous-mirror case you specified.
So I can’t state an admissible anchored mirror problem or predict its complexity without substituting an unnumbered algorithm or empirical finding for the required result.
The proponent names no anchor, and the paper supplies none of the kind this programme is meant to mirror. It has no numbered theorem, lemma, corollary, or proposition establishing a computational result. It proposes an EM procedure for approximate maximum-likelihood estimation, notes that it may get trapped in local optima, and reports simulation and user-study results. It gives no runtime or complexity classification, exact or approximation guarantee, or hardness result. So there is no paper result to continuize into a complexity question.
The obstacle remains if we give the mirror a better formulation. MOOCs and crowdsourcing can plainly have many participants, so multiplicity itself is not implausible. But the paper’s assessed outputs include the inferred ability and ranking of individual students. Grouping students by complete type can preserve that goal only when those students are indistinguishable in the observations and parameters the model uses; replacing individual rankings with an ability distribution or aggregate accuracy instead asks a different question. That aggregate inference problem might be worthwhile, but it would be a new computational question, not a mirror of a computational result in this paper.
This makes the negative case strong on the programme’s stated anchoring criterion, but weak as a universal claim that no useful continuous question could arise in any scenario. The paper does not establish such an impossibility; it simply offers no qualifying result for this mirror exercise.
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.