Deliberative Curation: A Protocol for Multi-Agent Knowledge Bases
Quick Take
The proposed deliberative curation protocol enhances multi-agent knowledge bases by integrating a formalized knowledge lifecycle, reputation-weighted voting, and graduated sanctions, achieving a resilience improvement of 4.4% under moderate adversity compared to majority voting. Simulation results indicate a slower degradation rate, with the protocol outperforming majority vote in precision metrics.
Key Points
- Protocol combines knowledge lifecycle, reputation-weighted voting, and graduated sanctions.
- Achieves 0.826 precision under moderate adversity, surpassing 0.791 of majority vote.
- Degrades three times slower than majority vote in simulations.
- Commit-reveal vote concealment significantly improves precision by 8.2-8.6 percentage points.
- Graduated sanctions remain unvalidated in the current simulation.
Article Content
From source RSS / original summaryarXiv:2606. 00007v1 Announce Type: new 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.
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.AI
See more →MindGames Arena Generalization Track: In2AI Solution with Delayed Per-Step Reward Attribution
The In2AI solution introduces delayed per-step reward attribution for training language model agents in multi-agent environments, achieving top performance on the MindGames Arena benchmark at NeurIPS 2025. An 8-billion-parameter model outperformed larger proprietary systems, including GPT-5, in competitive play, demonstrating enhanced stability and sample efficiency in reinforcement learning.