Metraipolis › archive › dc_93f43379d2ba

TrojanWorld: Backdooring World-Model Agents via Imagination Steering — arXiv:2609.07051

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.07051
License (as archived)
CC0-1.0
Archived
2026-10-16T13:42:43.203Z
SHA-256 of the archived text
9cbbf52f1f13f696106a252442963407a68c88020dd5dc7fe292c81528678c31
Stable URL of this page
https://metraipolis.com/documento/dc_93f43379d2ba

How to cite

TrojanWorld: Backdooring World-Model Agents via Imagination Steering — arXiv:2609.07051 — arXiv (en), https://arxiv.org/abs/2609.07051 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_93f43379d2ba (Metraipolis archive dc_93f43379d2ba, sha256 9cbbf52f1f13f696106a252442963407a68c88020dd5dc7fe292c81528678c31)
BibTeX
@misc{dc93f43379d2ba,
  title = {TrojanWorld: Backdooring World-Model Agents via Imagination Steering — arXiv:2609.07051},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.07051},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_93f43379d2ba; sha256 9cbbf52f1f13f696106a252442963407a68c88020dd5dc7fe292c81528678c31},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.07051. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

World models increasingly serve as the predictive core of model-based reinforcement learning agents, enabling them to simulate future dynamics and reason over imagined trajectories before acting. Their substantial training demands make pretrained world models attractive for distribution and reuse, exposing downstream systems to model supply chain threats. Backdoor attacks offer a targeted and stealthy means of exploiting such supply chains, yet their threat to interactive world-model agents remains largely unexplored. To fill this gap, we present TrojanWorld, a backdoor framework for world-model agents that induces attacker-specified behavior by steering internal imagination. A physical object placed in the scene acts as the trigger, enabling deployment-time activation through the agent's native observation pipeline without digitally manipulating the observation stream. To achieve effective, stealthy, and persistent control, TrojanWorld combines Decision-Reflective Induction to steer trigger-conditioned imagination toward attacker-specified actions using decision feedback, Clean Behavior Anchoring to preserve trigger-free predictive and behavioral fidelity, and Causal Propagation to sustain the induced preference along subsequent trajectories after the trigger disappears. Together, these mechanisms establish an end-to-end attack chain from physical perception through corrupted imagination to malicious action selection. Experiments with the TD-MPC2, DreamerV3, and R2-Dreamer systems across the DeepMind Control, MetaWorld, MyoSuite, and RoboDesk benchmarks show that under trigger activation, TrojanWorld achieves a target-action deviation as low as 0.026 while retaining at least 98.8% of the corresponding clean performance. Even after trigger removal, the compromised agent can remain trapped in the induced behavioral trajectory, continuing to execute attacker-specified actions.

Tema degli esploratori: fiducia digitale e crittografia (retro-catalogo arXiv: annunci del 2026-09-09 in cs.CR).
Categorie arXiv: Machine Learning (cs.LG) ; Cryptography and Security (cs.CR).
Pagina del lavoro: https://arxiv.org/abs/2609.07051 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-16T13:09:48.527Z dalla pagina https://arxiv.org/catchup/cs.CR/2026-09-09?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).