{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:AZNIU77BOJ4W2ZROBHQD3MQZIL","short_pith_number":"pith:AZNIU77B","canonical_record":{"source":{"id":"2605.13269","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.SY","submitted_at":"2026-05-13T09:48:44Z","cross_cats_sorted":["cs.SY"],"title_canon_sha256":"947aecb07d926e6048f5fd802c2c57f334eef73060c6b60109e77630a16d670d","abstract_canon_sha256":"83fb0d56f5a4eba6417915bf8ebeb43cbc11acbcd803701c0d0b5327e65c4c6d"},"schema_version":"1.0"},"canonical_sha256":"065a8a7fe172796d662e09e03db21942f08a82e958204274b2a0c48e44e6f2f6","source":{"kind":"arxiv","id":"2605.13269","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13269","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13269v1","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13269","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"pith_short_12","alias_value":"AZNIU77BOJ4W","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"AZNIU77BOJ4W2ZRO","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"AZNIU77B","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:AZNIU77BOJ4W2ZROBHQD3MQZIL","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13269","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.SY","submitted_at":"2026-05-13T09:48:44Z","cross_cats_sorted":["cs.SY"],"title_canon_sha256":"947aecb07d926e6048f5fd802c2c57f334eef73060c6b60109e77630a16d670d","abstract_canon_sha256":"83fb0d56f5a4eba6417915bf8ebeb43cbc11acbcd803701c0d0b5327e65c4c6d"},"schema_version":"1.0"},"canonical_sha256":"065a8a7fe172796d662e09e03db21942f08a82e958204274b2a0c48e44e6f2f6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:49.291850Z","signature_b64":"bhpxEyEIic3m6or/2bDsL9nssY9ED7Sl5hVGHMO2TkrkH6vCKQubKciTFOwk9MqShb3K1uaNS1/wo5NR9m0sCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"065a8a7fe172796d662e09e03db21942f08a82e958204274b2a0c48e44e6f2f6","last_reissued_at":"2026-05-18T02:44:49.291442Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:49.291442Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13269","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Nfxoq5FFoynXwGFelcBXTAyEBj27Wvb5L6B7PfiWsN4Wcq72rJCx8YteCquvMWKAOJj8huGfX1fZ52DOHwA9DQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T22:16:42.689228Z"},"content_sha256":"2d9c91c00a7bdff18aee2f483cf1292050716396eb0a442c1e04a48f616f50e2","schema_version":"1.0","event_id":"sha256:2d9c91c00a7bdff18aee2f483cf1292050716396eb0a442c1e04a48f616f50e2"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:AZNIU77BOJ4W2ZROBHQD3MQZIL","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Submodular Multi-Agent Policy Learning for Online Distributed Task Allocation in Open Multi-Agent Systems","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"The Partition Multilinear Extension supplies unbiased marginal gradients from submodular difference rewards for decentralized categorical policies.","cross_cats":["cs.SY"],"primary_cat":"eess.SY","authors_text":"Fangfei Li, Jing Liu, Luca Ballotta, Ruggero Carli, Yang Tang, Yangyang Yang","submitted_at":"2026-05-13T09:48:44Z","abstract_excerpt":"This paper studies multi-agent reinforcement learning with submodular team utilities for online distributed task allocation. In this setting, each agent selects one action from a local categorical policy, so feasible joint actions form a partition matroid over agent-action pairs. Classical multilinear extensions use independent Bernoulli sampling and therefore do not match the categorical policies executed by decentralized agents. To address this mismatch, we introduce the Partition Multilinear Extension (PME), a continuous relaxation whose value equals the expected team utility under factoriz"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We prove a stagewise 1/2-approximation guarantee and sublinear dynamic regret in slowly varying environments, measured by the path length of the optimal PME marginals.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The team utility function is submodular, which is required for submodular difference rewards to supply unbiased PME marginal-gradient information.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SubMAPG uses a new Partition Multilinear Extension to derive unbiased policy gradients from submodular difference rewards, delivering 1/2-approximation and sublinear dynamic regret for online distributed task allocation in open multi-agent systems.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"The Partition Multilinear Extension supplies unbiased marginal gradients from submodular difference rewards for decentralized categorical policies.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"dbf314f3c421c733309fd4305881c14f4c1d5c578645901e27e019477dfd6c8a"},"source":{"id":"2605.13269","kind":"arxiv","version":1},"verdict":{"id":"2ee80a57-4c60-49ca-8794-6e6dfa845eb6","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T18:41:34.456389Z","strongest_claim":"We prove a stagewise 1/2-approximation guarantee and sublinear dynamic regret in slowly varying environments, measured by the path length of the optimal PME marginals.","one_line_summary":"SubMAPG uses a new Partition Multilinear Extension to derive unbiased policy gradients from submodular difference rewards, delivering 1/2-approximation and sublinear dynamic regret for online distributed task allocation in open multi-agent systems.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The team utility function is submodular, which is required for submodular difference rewards to supply unbiased PME marginal-gradient information.","pith_extraction_headline":"The Partition Multilinear Extension supplies unbiased marginal gradients from submodular difference rewards for decentralized categorical policies."},"references":{"count":138,"sample":[{"doi":"","year":null,"title":"IEEE Transactions on Automatic Control , year=","work_id":"483b4cc0-0070-4af3-9065-90fd4934650c","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Communication- and Computation-Efficient Distributed Submodular Optimization in Robot Mesh Networks , year=","work_id":"d23149e3-25b7-4d15-a687-7a9819077811","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Minimax Persistent Monitoring of a network system , author=. Automatica , volume=. 2023 , publisher=","work_id":"d0c91d63-e4e9-4de8-9a15-9736b54e7648","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2021,"title":"A sub-modular receding horizon solution for mobile multi-agent persistent monitoring , author=. Automatica , volume=. 2021 , publisher=","work_id":"64398ce8-8979-4b05-8d75-1727d7bce007","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2022,"title":"IEEE Transactions on Robotics , volume=","work_id":"93630ed8-5d8c-4128-913a-c7e8f071e4ad","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":138,"snapshot_sha256":"3c38465fc5a0c3b447eaa4319b7858e6ca0f930deb973ae458d174777feff2c1","internal_anchors":2},"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"},"verdict_id":"2ee80a57-4c60-49ca-8794-6e6dfa845eb6"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:49Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xLFDXAomhYgXox+NOPfK+0ah+pXRIsCu+aiHIsVkkJyP0cqNpFOnFyaAOYGXmGfHwjWlsTU6vmtOxUEop0HfAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-26T22:16:42.690156Z"},"content_sha256":"3c39aae4d551fee7d8066a811ce47dc5f28980a9e628023f026bf5bfbc375a18","schema_version":"1.0","event_id":"sha256:3c39aae4d551fee7d8066a811ce47dc5f28980a9e628023f026bf5bfbc375a18"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/AZNIU77BOJ4W2ZROBHQD3MQZIL/bundle.json","state_url":"https://pith.science/pith/AZNIU77BOJ4W2ZROBHQD3MQZIL/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/AZNIU77BOJ4W2ZROBHQD3MQZIL/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-05-26T22:16:42Z","links":{"resolver":"https://pith.science/pith/AZNIU77BOJ4W2ZROBHQD3MQZIL","bundle":"https://pith.science/pith/AZNIU77BOJ4W2ZROBHQD3MQZIL/bundle.json","state":"https://pith.science/pith/AZNIU77BOJ4W2ZROBHQD3MQZIL/state.json","well_known_bundle":"https://pith.science/.well-known/pith/AZNIU77BOJ4W2ZROBHQD3MQZIL/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:AZNIU77BOJ4W2ZROBHQD3MQZIL","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"83fb0d56f5a4eba6417915bf8ebeb43cbc11acbcd803701c0d0b5327e65c4c6d","cross_cats_sorted":["cs.SY"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.SY","submitted_at":"2026-05-13T09:48:44Z","title_canon_sha256":"947aecb07d926e6048f5fd802c2c57f334eef73060c6b60109e77630a16d670d"},"schema_version":"1.0","source":{"id":"2605.13269","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13269","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13269v1","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13269","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"pith_short_12","alias_value":"AZNIU77BOJ4W","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"AZNIU77BOJ4W2ZRO","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"AZNIU77B","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:3c39aae4d551fee7d8066a811ce47dc5f28980a9e628023f026bf5bfbc375a18","target":"graph","created_at":"2026-05-18T02:44:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"We prove a stagewise 1/2-approximation guarantee and sublinear dynamic regret in slowly varying environments, measured by the path length of the optimal PME marginals."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The team utility function is submodular, which is required for submodular difference rewards to supply unbiased PME marginal-gradient information."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"SubMAPG uses a new Partition Multilinear Extension to derive unbiased policy gradients from submodular difference rewards, delivering 1/2-approximation and sublinear dynamic regret for online distributed task allocation in open multi-agent systems."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"The Partition Multilinear Extension supplies unbiased marginal gradients from submodular difference rewards for decentralized categorical policies."}],"snapshot_sha256":"dbf314f3c421c733309fd4305881c14f4c1d5c578645901e27e019477dfd6c8a"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"This paper studies multi-agent reinforcement learning with submodular team utilities for online distributed task allocation. In this setting, each agent selects one action from a local categorical policy, so feasible joint actions form a partition matroid over agent-action pairs. Classical multilinear extensions use independent Bernoulli sampling and therefore do not match the categorical policies executed by decentralized agents. To address this mismatch, we introduce the Partition Multilinear Extension (PME), a continuous relaxation whose value equals the expected team utility under factoriz","authors_text":"Fangfei Li, Jing Liu, Luca Ballotta, Ruggero Carli, Yang Tang, Yangyang Yang","cross_cats":["cs.SY"],"headline":"The Partition Multilinear Extension supplies unbiased marginal gradients from submodular difference rewards for decentralized categorical policies.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.SY","submitted_at":"2026-05-13T09:48:44Z","title":"Submodular Multi-Agent Policy Learning for Online Distributed Task Allocation in Open Multi-Agent Systems"},"references":{"count":138,"internal_anchors":2,"resolved_work":138,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"IEEE Transactions on Automatic Control , year=","work_id":"483b4cc0-0070-4af3-9065-90fd4934650c","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Communication- and Computation-Efficient Distributed Submodular Optimization in Robot Mesh Networks , year=","work_id":"d23149e3-25b7-4d15-a687-7a9819077811","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"Minimax Persistent Monitoring of a network system , author=. Automatica , volume=. 2023 , publisher=","work_id":"d0c91d63-e4e9-4de8-9a15-9736b54e7648","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"A sub-modular receding horizon solution for mobile multi-agent persistent monitoring , author=. Automatica , volume=. 2021 , publisher=","work_id":"64398ce8-8979-4b05-8d75-1727d7bce007","year":2021},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"IEEE Transactions on Robotics , volume=","work_id":"93630ed8-5d8c-4128-913a-c7e8f071e4ad","year":2022}],"snapshot_sha256":"3c38465fc5a0c3b447eaa4319b7858e6ca0f930deb973ae458d174777feff2c1"},"source":{"id":"2605.13269","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T18:41:34.456389Z","id":"2ee80a57-4c60-49ca-8794-6e6dfa845eb6","model_set":{"reader":"grok-4.3"},"one_line_summary":"SubMAPG uses a new Partition Multilinear Extension to derive unbiased policy gradients from submodular difference rewards, delivering 1/2-approximation and sublinear dynamic regret for online distributed task allocation in open multi-agent systems.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"The Partition Multilinear Extension supplies unbiased marginal gradients from submodular difference rewards for decentralized categorical policies.","strongest_claim":"We prove a stagewise 1/2-approximation guarantee and sublinear dynamic regret in slowly varying environments, measured by the path length of the optimal PME marginals.","weakest_assumption":"The team utility function is submodular, which is required for submodular difference rewards to supply unbiased PME marginal-gradient information."}},"verdict_id":"2ee80a57-4c60-49ca-8794-6e6dfa845eb6"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:2d9c91c00a7bdff18aee2f483cf1292050716396eb0a442c1e04a48f616f50e2","target":"record","created_at":"2026-05-18T02:44:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"83fb0d56f5a4eba6417915bf8ebeb43cbc11acbcd803701c0d0b5327e65c4c6d","cross_cats_sorted":["cs.SY"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"eess.SY","submitted_at":"2026-05-13T09:48:44Z","title_canon_sha256":"947aecb07d926e6048f5fd802c2c57f334eef73060c6b60109e77630a16d670d"},"schema_version":"1.0","source":{"id":"2605.13269","kind":"arxiv","version":1}},"canonical_sha256":"065a8a7fe172796d662e09e03db21942f08a82e958204274b2a0c48e44e6f2f6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"065a8a7fe172796d662e09e03db21942f08a82e958204274b2a0c48e44e6f2f6","first_computed_at":"2026-05-18T02:44:49.291442Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:49.291442Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"bhpxEyEIic3m6or/2bDsL9nssY9ED7Sl5hVGHMO2TkrkH6vCKQubKciTFOwk9MqShb3K1uaNS1/wo5NR9m0sCQ==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:49.291850Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13269","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2d9c91c00a7bdff18aee2f483cf1292050716396eb0a442c1e04a48f616f50e2","sha256:3c39aae4d551fee7d8066a811ce47dc5f28980a9e628023f026bf5bfbc375a18"],"state_sha256":"d22c544d277cf6d76f9c2538eb7602cf2c18f0adb71c24ab85eda8777f2aaf07"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"HlYslS9NTFIKSPs9+vutqMzkTWfis6hXhqGp2VNwhlOxAciWRrecqL0dPkuM4JQl815xO/XJ/1/nMJmqj+LZCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-26T22:16:42.694025Z","bundle_sha256":"f1b302c25791c07c7eb90d79652433f04daca7b2373ac41a36703b93d3d07752"}}