{"paper":{"title":"Zero-Shot Scalable Resilience in UAV Swarms: A Decentralized Imitation Learning Framework with Physics-Informed Graph Interactions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A policy trained only on 20-UAV swarms transfers directly to swarms of 500 UAVs without fine-tuning and outperforms baselines in reconnection reliability, recovery speed, motion safety, and efficiency.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Huan Lin, Lianghui Ding","submitted_at":"2026-04-17T07:06:42Z","abstract_excerpt":"Large-scale Unmanned Aerial Vehicle (UAV) failures can split an unmanned aerial vehicle swarm network into disconnected sub-networks, making decentralized recovery both urgent and difficult. Centralized recovery methods depend on global topology information and become communication-heavy after severe fragmentation. Decentralized heuristics and multi-agent reinforcement learning methods are easier to deploy, but their performance often degrades when the swarm scale and damage severity vary. We present Physics-informed Graph Adversarial Imitation Learning algorithm (PhyGAIL) that adopts centrali"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"A policy trained on 20-UAV swarms transfers directly to swarms of up to 500 UAVs without fine-tuning, and achieves better performance across reconnection reliability, recovery speed, motion safety, and runtime efficiency than representative baselines.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the bounded local interaction graphs combined with explicit attraction-repulsion message passing and scenario-adaptive imitation learning will generalize across arbitrary swarm sizes and fragmentation severities without overfitting to the 20-UAV training distribution.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"PhyGAIL uses bounded local graphs and physics-informed gated message passing in an imitation learning setup to achieve zero-shot scalable decentralized recovery in UAV swarms after failures.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A policy trained only on 20-UAV swarms transfers directly to swarms of 500 UAVs without fine-tuning and outperforms baselines in reconnection reliability, recovery speed, motion safety, and efficiency.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f567aee74689967c5d745d36d4ebc3da1d4514d650ae0aca65051dd04f6bc65d"},"source":{"id":"2604.15762","kind":"arxiv","version":2},"verdict":{"id":"f22baadc-7b54-4f61-93ca-2fec224096a9","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T09:29:36.801548Z","strongest_claim":"A policy trained on 20-UAV swarms transfers directly to swarms of up to 500 UAVs without fine-tuning, and achieves better performance across reconnection reliability, recovery speed, motion safety, and runtime efficiency than representative baselines.","one_line_summary":"PhyGAIL uses bounded local graphs and physics-informed gated message passing in an imitation learning setup to achieve zero-shot scalable decentralized recovery in UAV swarms after failures.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the bounded local interaction graphs combined with explicit attraction-repulsion message passing and scenario-adaptive imitation learning will generalize across arbitrary swarm sizes and fragmentation severities without overfitting to the 20-UAV training distribution.","pith_extraction_headline":"A policy trained only on 20-UAV swarms transfers directly to swarms of 500 UAVs without fine-tuning and outperforms baselines in reconnection reliability, recovery speed, motion safety, and efficiency."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.15762/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}