| paper | d4-conf-atal-2021-0026 |
| 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 names no theorem, lemma, proposition, or corollary asserting an algorithm, complexity bound, or hardness result; its reported percentage changes are simulation findings. Computing aggregate reach for a distribution of behavioral types is a plausible follow-on question, but it changes the graph model and has no named result here to mirror.
fails bit a — no named computational result to mirror
No eligible anchor: the paper's comparisons of personality, homophily, and fact-checking are simulation findings without a named computational result.
This paper has no eligible anchor. It contains no numbered theorem, lemma, corollary, or proposition asserting a computational complexity result. Its reported findings—such as the roughly \(12\%\) change in fake news going viral and the reduced apparent effectiveness of fact-checking when personality is included—come from simulation experiments. They are not named computational results under the requested standard. So there are no anchors to restate and no continuous problem to give *per anchor*.
The strongest plausible positive angle is an unanchored computational question about the paper’s central simulation: for a large platform where many users share a small number \(\tau\) of complete behavioral profiles, let \(\mu_t\) be the fraction of users of profile \(t\). Profiles would include every trait the model uses to determine behavior; homophilic exposure could be represented by a type- and degree-conditioned network kernel. Given news attributes, the sharing rules, a time horizon, and settings for personality, homophily, and fact-checking, ask for the expected aggregate reach of true and fake news—or the expected fraction of stories reaching a specified viral threshold. A solution would compute that quantity, exactly or within additive error \(\varepsilon\), and could compare the settings. A campaign or large platform with millions of users but relatively few repeated behavioral profiles is a plausible high-multiplicity regime.
That would ask for algorithms to compute spread in a population model, rather than for an equilibrium or an axiomatic property. The weakness is substantial: the paper’s agents occupy an individual network, where degree and local connections can shape viral spread, and its trait values are generated continuously. Replacing that graph with a type-level kernel is a meaningful modeling step, and the paper does not state a numbered complexity result that would anchor it. So this suggests a possible mirror, but it does not make an anchored case for one.
The proponent is right that this paper has no eligible anchor: it states no theorem, lemma, corollary, or proposition about the complexity of a computational problem. Its \(12\%\) viral-rate difference and fact-checking result are simulation findings, not complexity results. So there is no named result here to mirror.
The strongest reconstruction of the proposed mirror is: given masses \(\mu_t\) of complete behavioral types, a contact model, news attributes, and a time horizon, compute expected reach or the probability a story goes viral, perhaps comparing runs with and without personality or fact-checking. That is a coherent computational question, but it is a new one. In the paper’s sparse BA and CWS networks, type masses and even type-to-type contact rates need not determine spread: the arrangement of links and neighborhood correlations can change whether a cascade reaches a large audience. Encode those details in types, and the number of types can grow with the network, losing the high-multiplicity structure. Replace the network with independent mixing or a mean-field kernel, and the model no longer captures the topology and homophily whose effects the paper studies.
A stronger sparse-network limit, such as one that tracks local neighborhoods, might retain more of those effects. But that would be an algorithmic study of network diffusion, not a continuous mirror of a computational social-choice result. The paper defines no social-choice outcome, intervention optimization problem, or algorithmic task. Adding one—for example, choosing where to place fact-checks to change election winners—could make a worthwhile project, but it would add a new decision problem rather than continue this paper’s results.
That is the strongest negative case, and its limit is real: a carefully specified large-network diffusion problem could be useful in its own right. I cannot honestly rule out every such scenario. The case against a ChoCo mirror rests on the absence of an anchored computational result and on the fact that the plausible continuizations either lose the paper’s network structure or become a different network-science problem.
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.