AIG Research Publications

Four-phase pipeline that moves governance evaluation beyond accuracy metrics — attributing every governance decision to its upstream causal drivers using structural causal modeling (HillClimbSearch + PC) on 1,800 adversarial trial records. 11 high-confidence consensus causal edges. 99.89% correct rate across 6 scenario families.

Public-safe LUC-adjacent capability note describing AIG's governed RF analysis and edge experimentation substrate: provenance, synthetic fixtures, bounded modulation/FEC triage, GRAND-inspired lab experiments, and human-gated decoder orchestration.

Concise DICE-facing overview mapping AIG's 101/500 governed collective evidence to decentralized coordination, local inference control, contested environments, and human command authority.

Proposal-ready evidence narrative showing a validated path from a 101-record command catalog to 500 governed logical agents without uncontrolled all-to-all coordination.

Sanitized external-safe brief defining context governance as the control plane for agent collectives, with emphasis on interpretable context methods and bounded authority.

Architecture decision memo for the current edge/evaluation state bus, sparse coordination posture, and migration path toward sharded or distributed substrates.

Foundational thesis: alignment is not only a property of isolated models; it is a property of governance architecture, authority topology, and bounded delegation.

Formal systems specification for distributed governance cognitive architecture: roles, controls, trust boundaries, evaluation posture, and adversarial resilience.

Evidence record validating the current 101-record roster, fake-Qt preflight, and 500 logical-agent governed capacity harness with blocked/throttled interaction paths.