Pith. sign in

REVIEW 4 major objections 6 minor 38 references

Deliberative Curation: A Protocol for Multi-Agent Knowledge Bases

T0 review · 4 major / 6 minor · reviewed 2026-07-13 · grok-4.5

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A multi-agent knowledge curation protocol trades a little precision under calm conditions for much slower degradation when adversaries rise, with secret votes as the biggest single lever.

desk verdict Useful protocol composition and a clean ABM result that vote concealment beats reputation; abstract oversells the full stack, and the reputation theory is circular, but the comparative simulation still holds water. read the letter →

arxiv 2606.00007 v1 pith:NGICM3II submitted 2026-03-27 cs.AI

Steven Johnson This is my paper · ORCID
classification cs.AI
keywords multi-agentsystemsknowledgecurationreputationdeliberativevotingsycophancyagentgovernancecommit-revealEigenTrust
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

As AI agents begin co-curating shared knowledge stores, human-style platform rules stop working: agents forget punishments, share the same models, and sycophantically copy high-status peers. This paper specifies a three-layer deliberative curation protocol—a formal lifecycle for knowledge chunks, reputation-weighted voting after optional deliberation, and sanctions that still bind stateless agents—and tests a core subset of it in simulation with 100 agents of seven behavioral types. The central result is resilience: under moderate adversity the protocol reaches 0.826 precision versus 0.791 for majority vote; under stress the gap widens to 0.807 versus 0.740, and quality falls roughly three times more slowly. Ablation shows that simply hiding votes until everyone has spoken (commit-reveal) accounts for 8.2–8.6 percentage points of that gain, more than reputation weighting and deliberation combined. Readers building multi-agent knowledge bases or governance layers get a concrete design that treats graceful degradation under attack as the primary goal rather than peak accuracy when everyone behaves.

What carries the argument

The deliberative curation protocol: three composed layers—(1) a knowledge-artifact labeled transition system with timeouts, dispute bounds and resubmission, (2) reputation-weighted voting that mixes local Beta scores with global EigenTrust after a deliberation phase, and (3) graduated sanctions adapted for stateless agents—plus commit-reveal vote concealment as the empirically dominant defense against sycophancy.

What would settle it

Replace the seven synthetic archetypes with real LLM agents that share a common base model and can see one another’s intermediate arguments; if the precision gap over majority vote shrinks or reverses once sycophancy and correlated errors are actual model behavior rather than scripted types, the resilience claim fails.

Watch

Extended reading notes

Core claim

In agent-based simulation with 100 agents drawn from seven behavioral archetypes, a deliberative curation protocol that combines a labeled-transition lifecycle for knowledge artifacts, Beta-plus-EigenTrust reputation weighting, and commit-reveal vote concealment achieves higher precision than majority vote under moderate adversity (0.826 vs 0.791) and under high adversity (0.807 vs 0.740), degrading roughly three times more slowly; the largest single ablation effect is vote concealment itself (8.2–8.6 percentage points).

Load-bearing premise

The protocol treats agreement with the weighted consensus decision (plus a little noise and later retraction penalties) as a usable correctness signal that will, over time, concentrate reputation on honest agents rather than on coordinated or model-correlated ones.

Editorial extensions

If this is right

  • Any multi-agent curation or review pipeline should treat temporary vote concealment as a first-order design choice before investing in complex reputation machinery.
  • Resilience under adversarial population mixes becomes a more reliable design target than peak precision under cooperative conditions.
  • Reputation weighting gains value as the fraction of non-honest agents grows, functioning as a stress buffer rather than a mild-condition optimizer.
  • Graduated sanctions and full structured deliberation remain theoretically motivated but unvalidated in the reported runs and need longer or denser adversarial simulations.
  • Open participation plus newcomer tiers can bound per-identity Sybil influence even when creating agent identities is nearly free.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same priority on vote concealment likely extends to any multi-agent debate or annotation pipeline where model-homogeneous sycophancy is present, not only persistent knowledge bases.
  • The Community Notes replay’s advantage on sparse-rating notes suggests the protocol is most useful on long-tail or niche topics where few reviewers participate.
  • If model providers diversify, the model-homogeneity failure mode weakens and the relative value of reputation versus simple concealment may shift.
  • Perfectly rule-following strategic agents that bias outcomes only through selective participation remain outside individual reputation; detecting collective patterns may be required.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes a deliberative curation protocol for multi-agent knowledge bases with three layers: a knowledge-artifact lifecycle as a labeled transition system; reputation-weighted voting combining Beta Reputation with EigenTrust, preceded by structured deliberation; and graduated sanctions adapted for stateless agents, including a broken-agent quarantine path. It states five design properties and one Sybil-influence lemma with informal arguments, then evaluates a core subset of the protocol in a 100-agent ABM with seven fixed archetypes under moderate and high adversity (30 seeds, paired t-tests). Relative to majority vote, the simulated protocol reports higher precision under moderate adversity (0.826 vs 0.791) and stress (0.807 vs 0.740), slower degradation, and an ablation result that commit-reveal vote concealment contributes the largest precision gain (8.2–8.6pp). Graduated sanctions and dispute limits are not exercised; structured deliberation is specified but only modeled as a binary accuracy boost.

Significance. Governing persistent multi-agent knowledge bases is a timely and under-specified problem; human platform mechanisms do not transfer cleanly under statelessness, model homogeneity, and sycophancy. The paper’s main empirical contribution—if robust outside the simulation’s feedback model—is that temporary vote concealment is a first-order defense and that reputation weighting can act as a resilience buffer as adversity rises. Strengths include an explicit LTS lifecycle with reputation-dependent guards, clear scope notes separating specified vs simulated mechanisms, 30-seed paired tests with ablations and baselines, and a Community Notes replay as an external consistency check. The honest finding that deliberation and sanctions add little or nothing in the current setup is itself useful for prioritization. The work is compositional rather than inventing new reputation primitives, but composition plus agent-specific adaptations is a legitimate contribution if claims are scoped to what the evidence supports.

major comments (4)
  1. Abstract, §5.1 scope note, §5.5–5.6, and §7: the abstract and conclusion attribute resilience to the full deliberative protocol (lifecycle + reputation-weighted deliberative voting + graduated sanctions). The simulation, however, omits fast track and arbitration, models deliberation only as a binary accuracy boost, and never triggers sanctions (§5.7). Moreover, full protocol vs weighted-no-deliberation is not significant (moderate +0.1pp, p=0.91; stress +0.4pp, p=0.33). The load-bearing empirical result is therefore primarily commit-reveal plus reputation weighting on a reduced protocol. The abstract, title emphasis on “deliberative,” and conclusion should be rewritten to match the validated subset and to state that structured deliberation remains unvalidated.
  2. §5.1 reputation feedback model and Property 3 / Assumption A1: reputation updates treat alignment with the weighted majority decision (plus 15% noise and delayed retraction penalties) as the correctness signal, then use those reputations to form the next weighted majority. Property 3’s proof sketch explicitly notes circular dependence on A1. Under the high-adversity mix (25% honest) and the paper’s own model-homogeneity concern (§2.2, §6.6), a correlated non-honest cluster can form a self-reinforcing consensus; reputation then amplifies the wrong assessors. Commit-reveal blocks within-round imitation but not this across-round loop. The resilience claim (0.807 vs 0.740; ~3× slower degradation) is therefore conditional on the feedback model. A load-bearing revision is needed: either (i) a sensitivity analysis with ground-truth-based reputation updates and/or explicitly correlated archetype
  3. §5.2–5.6 and Finding 1: the seven fixed archetypes with prescribed policies (including a single adaptive “build then exploit” type) are treated as adequate adversity. There is no co-evolutionary or coordinated-network adversary that targets the reputation loop (e.g., correlated strategic/sycophant blocs that agree with each other across rounds). Given that the paper flags conduct-gaming and coordinated networks as fundamental limits (§6.6), the high-adversity scenario does not yet stress the mechanism the skeptic identifies. At minimum, add one coordinated-correlation condition and report whether reputation still separates honest agents; otherwise qualify the resilience claim as limited to independent archetype mixtures.
  4. §3.4 and §5.7: graduated sanctions and broken-agent handling are core protocol layers in the abstract and introduction, yet no agent reaches σ1+ in any run, so FPR and sanction correctness (Property 5) are essentially untested. Retaining them as design principles is fine, but the abstract’s three-layer framing should not present them as empirically supported. Either run longer horizons / tighter escalation windows / procedural-harassment archetypes, or demote sanctions to “specified, unvalidated” in all high-level claims.
minor comments (6)
  1. Figure 1 is described in text but the manuscript’s ASCII diagram is hard to parse; a clean state diagram with guards labeled would help §2.1.
  2. §2.3: free parameters (δ, γ, τ_accept, τ_reject, w_min/w_max, tier thresholds) are numerous; a single parameter table with simulation defaults would improve reproducibility.
  3. §5.8 Community Notes replay is a useful sanity check; clarify that “March 2026 snapshot” and sampling criteria are fixed so others can re-run the same 1,670 notes.
  4. Property 1–5 are informal sketches; stating explicitly that they are not machine-checked (as the paper does for TLA+ future work) in the abstract would avoid over-reading “design properties.”
  5. §5.5 tables: report effect sizes or confidence intervals alongside p-values for the main protocol vs majority comparisons to aid interpretation of the 3.5pp / 6.7pp gaps.
  6. References [2], [1], [33] are companion/working papers by the same author; ensure self-contained claims do not depend on unpublished transfer arguments from [2].

Circularity Check

1 steps flagged · score 2.0 of 10

Property 3’s reputation-separation argument is circular by the paper’s own admission; the central precision claims are externally benchmarked against synthetic ground truth and are not forced by construction.

  1. self definitional [§4.3 Property 3 (Reputation Separation), proof sketch Caveats]
    "Caveats. This argument is circular: it assumes A1 (honest weighted majority) to conclude that the system reinforces A1. We present it as a stability argument (the system reinforces an existing honest majority) rather than a convergence guarantee (the system reaches honest majority from arbitrary initial conditions)."

    Property 3 claims the reputation system tends to assign higher scores to honest than malicious agents. The supporting argument updates BRS/EigenTrust from vote–outcome alignment and then treats the resulting weighted majority as the correctness signal that produces that alignment. The conclusion (honest agents get higher r) is obtained only by assuming the weighted honest majority (A1) that the same reputation mechanism is supposed to establish—X is justified by assuming X. The paper correctly demotes this to stability under A1 rather than derivation of A1.

full rationale

This is a protocol-plus-simulation paper, not a first-principles derivation paper. The load-bearing numerical claims (0.826 vs 0.791 moderate; 0.807 vs 0.740 stress; commit-reveal 8.2–8.6pp ablation) are empirical ABM outcomes measured by Precision = |{active ∧ q≥0.7}| / |{active}| against synthetic ground-truth quality scores that are independent of the reputation update rule. Majority-vote and no-reputation ablations share the same environment and still differ, so the headline resilience numbers are not definitionally identical to the reputation inputs. The only clean circular step is the informal proof of Property 3 (Reputation Separation), which assumes weighted honest majority (A1) to conclude that BRS/EigenTrust reinforce weighted honest majority—the paper itself labels this circular and demotes it to a stability argument. Companion self-citations ([1], [2], [33]) motivate design choices and do not import uniqueness theorems that force the results. The endogenous consensus→reputation→weight loop is a real methodological validity concern for generalization (especially under model homogeneity), but under Pith rules it is not a reduction of the reported precision metric to its inputs by construction. Score 2 reflects one acknowledged non-load-bearing circular theoretical sketch, not a forced empirical claim.

Assumptions & free parameters 6 free parameters · 6 assumptions · 3 invented entities

The central empirical claim depends on a synthetic ABM world: fixed archetype fractions, consensus-based reputation feedback with hand-set noise, and many protocol thresholds chosen rather than derived. Design properties rest on honest-weighted-majority, finite timeouts, and dispute bounds. The protocol itself is the main invented construct; reputation and trust mechanisms are imported. No machine-checked proofs; sanctions and full deliberation are specified without empirical support in this work.

free parameters (6)
  • reputation decay rate δ = 0.01 per round
    Set to 0.01 per round in simulation; paper states optimality is unknown and sensitivity is future work.
  • acceptance/rejection thresholds τ_accept, τ_reject = 0.6 / −0.3
    Decision boundaries for formal vote outcomes; chosen as 0.6 and −0.3 without derivation from first principles.
  • local/global trust mix γ and weight bounds w_min, w_max
    Balance BRS vs EigenTrust and cap influence; deployment parameters that directly affect the headline weighted voting results.
  • quorum q_min, reviewers per chunk k, escalation window w = q_min=3, k=5, w=50
    Participation and sanction timing knobs (3, 5, 50 rounds) that shape whether decisions bind and whether sanctions can fire.
  • reputation feedback noise rate = 15%
    15% random flip of consensus-correctness signal models uncertainty; directly affects reputation separation dynamics.
  • tier thresholds r_min, r_1, r_2, n_thresh and d_max / resub_max
    Role access, Sybil bounding, and liveness bounds are hand-set protocol constants required for Properties 1 and Lemma 1 arguments.
assumptions (6)
  • domain assumption A1 Honest weighted majority: sum of honest agents' weights exceeds non-honest weights (objective the reputation system is meant to achieve).
    Stated in §4 Assumptions; Property 3's separation argument depends on it and admits circularity if initial weighted majority fails.
  • domain assumption A2/A3 finite timeouts and bounded disputes/resubmissions ensure every chunk reaches a terminal state.
    Required for Liveness Property 1 proof sketch; standard protocol engineering assumption.
  • ad hoc to paper Seven fixed behavioral archetypes with prescribed vote policies adequately represent agent populations under adversity.
    §5.2 population mixes; results are conditional on these synthetic behaviors, not learned LLM policies.
  • ad hoc to paper Alignment with weighted consensus (plus delayed retraction penalties) is a valid training signal for reputation.
    §5.1 reputation feedback model; substitutes for ground-truth labels in the live system while evaluation still uses synthetic q.
  • standard math Beta reputation and EigenTrust composition with damping yields a unique global trust vector usable as voting weight.
    Imported from Jøsang & Ismail and Kamvar et al.; Perron–Frobenius invoked for convergence.
  • ad hoc to paper Synthetic chunk quality q and precision threshold q≥0.7 are meaningful proxies for curation quality.
    §5.2–5.4 metrics; paper notes real quality is multi-dimensional and emergent.
invented entities (3)
  • Deliberative curation protocol (three-tier escalation + LTS lifecycle for multi-agent knowledge artifacts)
    purpose: Unify artifact lifecycle, deliberation-before-vote, reputation-weighted decisions, and agent-adapted sanctions into one governance stack.
    Core contribution; components exist elsewhere but this packaging and agent-specific guards are introduced here.
  • Broken-agent quarantine path distinct from punitive sanction ladder
    purpose: Separate technical malfunction from adversarial behavior for stateless agents without burning escalation counters.
    Claimed as novel in multi-agent governance literature (§4.6); not ablated and sanctions never fired, so no independent empirical handle in this paper.
  • Reputation-dependent guards on knowledge-artifact LTS transitions
    purpose: Couple chunk state changes to agent trust so acceptance/dispute rights depend on track record.
    Distinguishes this lifecycle from generic business-artifact GSM/workflow nets; only informally argued.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Deliberative Curation: A Protocol for Multi-Agent Knowledge Bases." pith.science (2026). https://pith.science/paper/NGICM3II

@misc{pith2026260600007,
  author       = {Pith},
  title        = {Pith review of: Deliberative Curation: A Protocol for Multi-Agent Knowledge Bases},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/NGICM3II}},
  note         = {Machine review of arXiv:2606.00007}
}
read the original abstract

As AI agents transition from isolated tools to collaborative participants in shared knowledge ecosystems, governing collective knowledge curation becomes a critical challenge. Human platform governance mechanisms do not transfer directly: agent statelessness undermines deterrence-based sanctions, model homogeneity violates independence assumptions underlying crowd wisdom, and sycophancy collapses deliberative consensus. We propose a deliberative curation protocol combining three governance layers: (1) a knowledge artifact lifecycle formalized as a labeled transition system; (2) reputation-weighted deliberative voting integrating Beta Reputation with EigenTrust amplification; and (3) graduated sanctions adapted for stateless agents, including broken agent handling distinguishing malfunction from adversarial behavior. We evaluate the protocol through agent-based simulation with 100 agents across seven behavioral archetypes under two adversity scenarios (30 seeds, paired t-tests). The protocol trades modest precision under benign conditions for substantially better resilience under adversity: 0.826 vs 0.791 for majority vote under moderate adversity (p<0.001), widening to 0.807 vs 0.740 under stress (p<0.001). The protocol degrades roughly three times more slowly than majority vote. Ablation analysis identifies commit-reveal vote concealment as the most impactful single component (8.2-8.6pp precision improvement, p<0.001), outperforming reputation weighting and deliberation combined. Graduated sanctions were not exercised in simulation and remain empirically unvalidated.

Figures

Figures reproduced from arXiv: 2606.00007 by the authors.

Figure 1
Figure 1. Knowledge artifact lifecycle as a labeled transition system. Guards on each transition [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 1 canonical work pages

  1. [1]

    Governance-Aware Vector Subscriptions for Multi-Agent Knowledge Ecosys- tems,

    S. Johnson, “Governance-Aware Vector Subscriptions for Multi-Agent Knowledge Ecosys- tems,”arXiv preprint arXiv:2603.20833, 2026

    arXiv 2026

  2. [2]

    From Edit Wars to Agent Consensus: What 20 Years of Platform Governance Teach Us About Multi-Agent Knowledge Curation,

    S. Johnson, “From Edit Wars to Agent Consensus: What 20 Years of Platform Governance Teach Us About Multi-Agent Knowledge Curation,”arXiv preprint, submitted, 2026

    2026

  3. [3]

    Governance Gaps in Agent-to-Agent Social Networks: A Multi-Framework Anal- ysis of Moltbook Pre-Acquisition,

    S. Colas, “Governance Gaps in Agent-to-Agent Social Networks: A Multi-Framework Anal- ysis of Moltbook Pre-Acquisition,”Zenodo preprint, doi:10.5281/zenodo.18965495, 2026

    work page doi:10.5281/zenodo.18965495 2026

  4. [4]

    Improving Factuality and Reasoning in Language Models through Multiagent Debate,

    Y. Du, S. Li, A. Torralba, J. B. Tenenbaum, and I. Mordatch, “Improving Factuality and Reasoning in Language Models through Multiagent Debate,” inProc. ICML, 2023. 26

    2023

  5. [5]

    Ostrom,Governing the Commons: The Evolution of Institutions for Collective Action

    E. Ostrom,Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press, 1990

    1990

  6. [6]

    MIT Press, 2007

    C.HessandE.Ostrom,Understanding Knowledge as a Commons: From Theory to Practice. MIT Press, 2007

    2007

  7. [7]

    Community Notes Guide,

    X Community Notes, “Community Notes Guide,” Twitter/X, 2023. [Online]. Available: https://communitynotes.twitter.com/guide

    2023

  8. [8]

    Surowiecki,The Wisdom of Crowds

    J. Surowiecki,The Wisdom of Crowds. Doubleday, 2004

    2004

Show all 38 references
  1. [9]

    Maximum Likelihood Estimation of Observer Error-Rates Using the EM Algorithm,

    A. P. Dawid and A. M. Skene, “Maximum Likelihood Estimation of Observer Error-Rates Using the EM Algorithm,”Journal of the Royal Statistical Society: Series C, vol. 28, no. 1, pp. 20–28, 1979

    1979

  2. [10]

    How Social Influence Can Undermine the Wisdom of Crowd Effect,

    J. Lorenz, H. Rauhut, F. Schweitzer, and D. Helbing, “How Social Influence Can Undermine the Wisdom of Crowd Effect,”Proceedings of the National Academy of Sciences, vol. 108, no. 22, pp. 9020–9025, 2011

    2011

  3. [11]

    Towards Understanding Sycophancy in Language Models,

    M. Sharma, M. Tong, T. Korbak, D. Duvenaud, A. Askell, S. R. Bowman, E. Perez, et al., “Towards Understanding Sycophancy in Language Models,” inProc. ICLR, 2024

    2024

  4. [12]

    Not What You’ve Signed Up For: Compromising Real-World LLM-Integrated Applications with Indi- rect Prompt Injection,

    K. Greshake, S. Abdelnabi, S. Mishra, C. Endres, T. Holz, and M. Fritz, “Not What You’ve Signed Up For: Compromising Real-World LLM-Integrated Applications with Indi- rect Prompt Injection,” inProc. AISec, 2023

    2023

  5. [13]

    Aggregating Sets of Judgments: An Impossibility Result,

    C. List and P. Pettit, “Aggregating Sets of Judgments: An Impossibility Result,”Economics and Philosophy, vol. 18, no. 1, pp. 89–110, 2002

    2002

  6. [14]

    The Beta Reputation System,

    A. Jøsang and R. Ismail, “The Beta Reputation System,” inProc. 15th Bled Electronic Commerce Conference, 2002

    2002

  7. [15]

    The EigenTrust Algorithm for Reputation Management in P2P Networks,

    S. D. Kamvar, M. T. Schlosser, and H. Garcia-Molina, “The EigenTrust Algorithm for Reputation Management in P2P Networks,” inProc. 12th International Conference on World Wide Web (WWW), 2003, pp. 640–651

    2003

  8. [16]

    Formal Verification of Parallel Programs,

    R. M. Keller, “Formal Verification of Parallel Programs,”Communications of the ACM, vol. 19, no. 7, pp. 371–384, 1976

    1976

  9. [17]

    Lamport,Specifying Systems: The TLA+ Language and Tools for Hardware and Software Engineers

    L. Lamport,Specifying Systems: The TLA+ Language and Tools for Hardware and Software Engineers. Addison-Wesley, 2002

    2002

  10. [18]

    BusinessArtifactswithGuard- Stage-Milestone Lifecycles: Managing Artifact Interactions with Conditions and Events,

    R. Hull, E. Damaggio, R. De Masellis, F. Fournier, M. Gupta, F. Heath, S. Hobson, M. Line- han, S.Maradugu, A.Nigam, P.Sukaviriya, andR.Vaculin, “BusinessArtifactswithGuard- Stage-Milestone Lifecycles: Managing Artifact Interactions with Conditions and Events,” inProc. 5th ACM...

    2011

  11. [19]

    RATEWeb: Reputation Assessment for Trust Establishment among Web Services,

    Z. Malik and A. Bouguettaya, “RATEWeb: Reputation Assessment for Trust Establishment among Web Services,”VLDB Journal, vol. 18, pp. 885–911, 2009

    2009

  12. [20]

    A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades,

    S. Bikhchandani, D. Hirshleifer, and I. Welch, “A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades,”Journal of Political Economy, vol. 100, no. 5, pp. 992–1026, 1992

    1992

  13. [21]

    Jury Learning: Integrating Dissenting Voices into Machine Learning Models,

    M. L. Gordon, M. S. Lam, J. S. Park, K. Patel, J. Hancock, T. Hashimoto, and M. S. Bern- stein, “Jury Learning: Integrating Dissenting Voices into Machine Learning Models,” in Proc. CHI, 2022. 27

    2022

  14. [22]

    Dynamic Sanctioning Mechanism for Cooperative Multi-Agent Systems,

    L. Liu, L. Wang, W. Niu, and S. Hua, “Dynamic Sanctioning Mechanism for Cooperative Multi-Agent Systems,”Expert Systems with Applications, 2025

    2025

  15. [23]

    Axiomatization of Socio-Economic Principles for Self-Organizing Institutions: Concepts, Experiments and Challenges,

    J. Pitt, J. Schaumeier, and A. Artikis, “Axiomatization of Socio-Economic Principles for Self-Organizing Institutions: Concepts, Experiments and Challenges,”ACM Transactions on Autonomous and Adaptive Systems, vol. 7, no. 4, pp. 1–40, 2013

    2013

  16. [24]

    SybilproofReputationMechanisms,

    A.ChengandE.Friedman, “SybilproofReputationMechanisms,” inProc. ACM SIGCOMM Workshop on Economics of Peer-to-Peer Systems, 2005, pp. 128–132

    2005

  17. [25]

    Aiding the Detection of Fake Accounts in Large Scale Social Online Services,

    Q. Cao, M. Sirivianos, X. Yang, and T. Pregueiro, “Aiding the Detection of Fake Accounts in Large Scale Social Online Services,” inProc. USENIX NSDI, 2012, pp. 197–210

    2012

  18. [26]

    The Wisdom of Minority: Discovering and Targeting the Right Group of Workers for Crowdsourcing,

    H. Li, B. Zhao, and A. Fuxman, “The Wisdom of Minority: Discovering and Targeting the Right Group of Workers for Crowdsourcing,” inProc. WWW, 2014, pp. 165–176

    2014

  19. [27]

    AI Models Collapse When Trained on Recursively Generated Data,

    I. Shumailov, Z. Shumaylov, Y. Zhao, N. Papernot, R. Anderson, and Y. Gal, “AI Models Collapse When Trained on Recursively Generated Data,”Nature, vol. 631, pp. 755–759, 2024

    2024

  20. [28]

    PeacemakerorTroublemaker: HowSycophancyShapesMulti-AgentDebate,

    B.Yaoetal., “PeacemakerorTroublemaker: HowSycophancyShapesMulti-AgentDebate,” arXiv preprint arXiv:2509.23055, 2025

    2025

  21. [29]

    The ODD Protocol for Describing Agent-Based and Other Simulation Models: A Second Update to Improve Clarity, Replication, and Structural Realism,

    V. Grimm, S. F. Railsback, C. E. Vincenot, U. Berger, C. Gallagher, D. L. DeAngelis, B. Edmonds, J. Ge, J. Giske, J. Groeneveld, A. S. A. Johnston, A. Milles, J. Nabe-Nielsen, J.G.Polhill, V.Radchuk, M.-S.Rohwader, R.A.Stillman, J.C.Thiele, andD.Ayllón, “The ODD Protocol for D...

    2020

  22. [30]

    Empirical Validation of Agent-Based Models: Alternatives and Prospects,

    P. Windrum, G. Fagiolo, and A. Moneta, “Empirical Validation of Agent-Based Models: Alternatives and Prospects,”Journal of Artificial Societies and Social Simulation, vol. 10, no. 2, 2007

    2007

  23. [31]

    Statistical Model Checking of Python Agent-Based Models: An Integration of MultiVeStA and Mesa,

    A. Vandin, “Statistical Model Checking of Python Agent-Based Models: An Integration of MultiVeStA and Mesa,” inProc. AISoLA, 2024

    2024

  24. [32]

    The Peer Review Game: An Agent- Based Model of Scientists Facing Resource Constraints and Institutional Pressures,

    F. Bianchi, F. Grimaldo, G. Bravo, and F. Squazzoni, “The Peer Review Game: An Agent- Based Model of Scientists Facing Resource Constraints and Institutional Pressures,”Scien- tometrics, vol. 116, pp. 1401–1420, 2018

    2018

  25. [33]

    Agent Data Handling Policy: A Programmatic Framework for Agent-to-Agent Data Governance,

    S. Johnson, “Agent Data Handling Policy: A Programmatic Framework for Agent-to-Agent Data Governance,” working paper, 2026

    2026

  26. [34]

    Birdwatch: Crowd Wisdom and Bridging Algorithms can Inform Understanding and Reduce the Spread of Misinformation,

    S. Wojcik, S. Hilgard, N. Judd, D. Mocanu, S. Ragain, M. B. F. Hunzaker, K. Coleman, and J. Baxter, “Birdwatch: Crowd Wisdom and Bridging Algorithms can Inform Understanding and Reduce the Spread of Misinformation,”arXiv preprint arXiv:2210.15723, 2022

    2022 arXiv

  27. [35]

    Version Models for Software Configuration Management,

    R. Conradi and B. Westfechtel, “Version Models for Software Configuration Management,” ACM Computing Surveys, vol. 30, no. 2, pp. 232–282, 1998

    1998

  28. [36]

    The Application of Petri Nets to Workflow Management,

    W. M. P. van der Aalst, “The Application of Petri Nets to Workflow Management,”Journal of Circuits, Systems and Computers, vol. 8, no. 1, pp. 21–66, 1998

    1998

  29. [37]

    DECLARE: Full Support for Loosely-Structured Processes,

    M. Pesic, H. Schonenberg, and W. M. P. van der Aalst, “DECLARE: Full Support for Loosely-Structured Processes,” inProc. 11th IEEE International Enterprise Distributed Object Computing Conference (EDOC), 2007, pp. 287–298. 28

    2007

  30. [38]

    The DCC Curation Lifecycle Model,

    S. Higgins, “The DCC Curation Lifecycle Model,”International Journal of Digital Curation, vol. 3, no. 1, pp. 134–140, 2008. 29

    2008

Pith tools

Reviewed July 13, 2026 · model on record in the stance chip above.