Multi-Aspect Runtime Verification for Simulation-Based V&V of LLM-Enabled Autonomous Agents

Key points

• LLM-based agents are entering decision-support roles in defence staff work, where the obligations they must respect are already written down and binding, and where retraining is not available as a control because models arrive as procured components.1
• What can be placed under engineering control is the interface between the agent and the systems it acts on.1
• Those obligations are at once spatial, temporal and text-semantic, and a violation typically lives in the composition of a multi-step interaction, which is why per-event guardrails miss sequential tool-attack chains.1
• The spatial aspect is interpreted over a weighted two-sorted location graph in which mission geometry and information-release topology are one object; we show that these spatial obligations are not in general subsumed by a first-order temporal specification.1

Sources (1)

• [1]Multi-Aspect Runtime Verification for Simulation-Based V&V of LLM-Enabled Autonomous Agents
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 6, 06:00 PM

  We present a multi-aspect runtime-verification framework that decomposes a natural-language policy clause into a typed spatial/temporal/semantic triple over one canonical event stream, checks each aspect with its own monitoring specification, and fuses the verdicts through a four-valued algebra that carries provenance.

  LLM-based agents are entering decision-support roles in defence staff work, where the obligations they must respect are already written down and binding, and where retraining is not available as a control because models arrive as procured components.

Extractive summary: sentences quoted from the sources.