Search papers, labs, and topics across Lattice.
This paper introduces MELD, a novel protocol designed to enable autonomous agents to reconcile and share knowledge across distributed memories without losing information due to conflicting claims. By employing a five-outcome procedure for incoming claims and leveraging a knowledge graph as its runtime model, MELD ensures that agents can maintain coherence in their knowledge while preserving contradictions for future adjudication. The results demonstrate that MELD achieves recall performance comparable to centralized systems while significantly reducing storage requirements and improving message efficiency in distributed environments.
MELD allows autonomous agents to reconcile conflicting knowledge without losing information, achieving superior recall and efficiency compared to traditional methods.
Autonomous agents share a transport and can call each other's tools, but they cannot share what they know: no protocol lets two agents'memories reconcile a fact phrased two ways, link related facts held apart, or reconcile contradictory knowledge without silently discarding either claim. We present MELD, a self-managing coherence mechanism for a federation of agent memories whose run-time model is the knowledge graph itself. Each brain admits every incoming claim through a five-outcome procedure (insert, merge, relate, conflict, or reject), decided from three signals (scoped claim-key identity, embedding similarity, and a natural-language-inference verdict) under context and freshness gates, and acting through exactly one auditable, authenticated Patch, the only object that mutates state. A binding onto standard publish/subscribe transport with a per-claim status CRDT keeps sovereign brains coherent in claim status without a coordinator: self-healing after partitions and under lossy routing, and self-protecting against silent rewrite by a peer, under a benign-fault model. MELD does not adjudicate truth; a detected contradiction is preserved for later adjudication, never silently resolved. On HotpotQA distractor, distributed merge is recall-non-inferior to a centralized store under a pre-specified equivalence test and recall-superior to naive union at about 11% less live storage; the merge classifier separates at AUC 0.968 with a 0.013 false-merge rate on adjudicated candidate pairs; the status CRDT reconverges in 30/30 real partition-heal trials where last-writer-wins manages 11/30; and semantic routing delivers about 3x fewer messages at matched recall. We evaluate on a real computing continuum spanning an operator-grade 5G edge, national HPC, and a local tier, with empirically calibrated thresholds.
