Driving · 2026-11-03

GURU: A Hierarchical Neuro-Symbolic Multi-Agent System for Verification-Aware Automation of Mission-Critical Digital Engineering and Operations

Allan Grosvenor, Anton Zemlyansky, Abdul Wahab, Kyrylo Bohachov, Aras Dogan, Ryland Adams, Dwyer Deighan

arXiv:2611.01001PDF

AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.

TL;DR

GURU is a hierarchical neuro-symbolic multi-agent system for verification-aware automation of mission-critical digital engineering and operations. Learned agents propose, rank, and accelerate actions, while a symbolic rules engine checks consequential actions against hard limits before execution; validated physics confirms and decides in a closed control loop using a shared internal language. Reported results include CFD speedups and higher ParaView control success, and improved conjunction-risk performance for a frozen self-supervised world model.

Why it matters

The authors emphasize a partition of authority so that models propose and accelerate, but do not decide what is safe. By adding symbolic limit checks and validated physics, GURU targets trustworthy automation in nuclear/aviation/national-security contexts and demonstrates performance gains on engineering workflows while keeping safety constraints enforced.

Method

  • Hierarchical neuro-symbolic multi-agent setup with one partition of authority: learned models propose/rank/accelerate; a symbolic rules engine verifies every consequential action against hard limits before execution; validated physics confirms and decides.
  • Runs a workflow as a closed control loop with Navigation, Action, and Planning agents sharing one Common Internal Language and one training factory.
  • Uses a harnessed language model while placing a validated reference at the model’s center; includes a frozen self-supervised world model where language models cannot be trusted to decide.
Abstract (from arXiv)

An AI can find a safe action faster than any human, but it must never decide what counts as safe. We present GURU, a hierarchical neuro-symbolic multi-agent system for mission-critical digital engineering and operations, built on one partition of authority: learned models propose, rank, and accelerate; a symbolic rules engine checks every consequential action against hard limits before it executes; and validated physics confirms and decides. GURU runs a workflow as a closed control loop, with Navigation, Action, and Planning agents sharing one Common Internal Language and one training factory. Nuclear, aviation, and national-security guidance now requires this pattern; a harnessed language model keeps the model at the center, GURU puts a validated reference there. In computational fluid dynamics, the workflow raised the converged-simulation rate on unseen aircraft geometries from under 1 percent to 74 percent over 100 single-attempt cases, ran 54 live cases to Mach 6 on three facilities, and an ablation put the symbolic layer at 13 points of boundary-layer capture. Knowledge-graph retrieval and role adapters raised ParaView control success from 55 to 75 percent. In space domain awareness, where a language model cannot be trusted to decide, a frozen self-supervised world model under validated propagation scored highest of five methods at conjunction risk, including fully fine-tuned language models, with several-times-lower seed variance: the trustworthy design gives up nothing in accuracy, and the recipe transferred to reactor anomaly detection. A radar-to-avoidance debris pipeline was demonstrated in stages on a 2026 reentry; the hierarchy coordinates simulated teams of 200 aircraft. Training scaled near-linearly past 1,000 Aurora nodes and ran at 8,000. The approach extends to a control-grade microreactor digital twin with validated physics, not a learned model, as the safety basis.

Related papers

Publication record

Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.

First announced by arXiv: 2026-11-03 · Indexed by Pub-AI: 2026-11-03 UTC

preprint

GURU: A Hierarchical Neuro-Symbolic Multi-Agent System for Verification-Aware Automation of Mission-Critical Digital Engineering and Operations

Allan Grosvenor, Anton Zemlyansky, Abdul Wahab, Kyrylo Bohachov, Aras Dogan, Ryland Adams, Dwyer Deighan

First announced by arXiv: 2026-11-03 · Original record ↗

arxiv
2611.01001
Explicit source versions (1)
  • v1 · Date not supplied
  1. Preprint available

    Source date: 2026-11-03 · First detected: 2026-11-03 UTC · Historical baseline

    Source evidence (1)
    • Official evidence ↗ · RSS guid=oai:arXiv.org:2611.01001v1; arxiv:announce_type=new; item pubDate (official metadata)
Abstract · arxiv

An AI can find a safe action faster than any human, but it must never decide what counts as safe. We present GURU, a hierarchical neuro-symbolic multi-agent system for mission-critical digital engineering and operations, built on one partition of authority: learned models propose, rank, and accelerate; a symbolic rules engine checks every consequential action against hard limits before it executes; and validated physics confirms and decides. GURU runs a workflow as a closed control loop, with Navigation, Action, and Planning agents sharing one Common Internal Language and one training factory. Nuclear, aviation, and national-security guidance now requires this pattern; a harnessed language model keeps the model at the center, GURU puts a validated reference there. In computational fluid dynamics, the workflow raised the converged-simulation rate on unseen aircraft geometries from under 1 percent to 74 percent over 100 single-attempt cases, ran 54 live cases to Mach 6 on three facilities, and an ablation put the symbolic layer at 13 points of boundary-layer capture. Knowledge-graph retrieval and role adapters raised ParaView control success from 55 to 75 percent. In space domain awareness, where a language model cannot be trusted to decide, a frozen self-supervised world model under validated propagation scored highest of five methods at conjunction risk, including fully fine-tuned language models, with several-times-lower seed variance: the trustworthy design gives up nothing in accuracy, and the recipe transferred to reactor anomaly detection. A radar-to-avoidance debris pipeline was demonstrated in stages on a 2026 reentry; the hierarchy coordinates simulated teams of 200 aircraft. Training scaled near-linearly past 1,000 Aurora nodes and ran at 8,000. The approach extends to a control-grade microreactor digital twin with validated physics, not a learned model, as the safety basis.