{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:N3K3BDIEGEX2C7MFBOZL3G2WHN","short_pith_number":"pith:N3K3BDIE","schema_version":"1.0","canonical_sha256":"6ed5b08d04312fa17d850bb2bd9b563b6d1fbe4cc18669d34b0da23f0d68d137","source":{"kind":"arxiv","id":"2604.12645","version":2},"attestation_state":"computed","paper":{"title":"Contextual Multi-Task Reinforcement Learning for Autonomous Reef Monitoring","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Contextual multi-task reinforcement learning trains one policy to handle multiple reef monitoring tasks.","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Frank Kirchner, Mariela De Lucas Alvarez, Melvin Laux, Rebecca Adam, Rina Alo, S\\\"oren T\\\"opper, Yi-Ling Liu","submitted_at":"2026-04-14T12:16:56Z","abstract_excerpt":"Although autonomous underwater vehicles promise the capability of marine ecosystem monitoring, their deployment is fundamentally limited by the difficulty of controlling vehicles under highly uncertain and non-stationary underwater dynamics. To address these challenges, we employ a data-driven reinforcement learning approach to compensate for unknown dynamics and task variations. Traditional single-task reinforcement learning has a tendency to overfit the training environment, thus, limit the long-term usefulness of the learnt policy. Hence, we propose to use a contextual multi-task reinforcem"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2604.12645","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.RO","submitted_at":"2026-04-14T12:16:56Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"142d5ac54bb02a1c6b6be8ec8c83df63fd0e7cd2f523964f178729e67a20a1af","abstract_canon_sha256":"bb3c172d3b226d9c19435e42539de97835efdc34c3ee87ef09baa4840e3b5d4e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-04T01:08:50.037737Z","signature_b64":"fRJdSXmbwptTF2RvY8JnJ3RYFMygWpjvDhWTWTDDuaQHyTYszTjMKMo/HZYc2zLJzeCwNNKsiL8zI07AxIKoBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6ed5b08d04312fa17d850bb2bd9b563b6d1fbe4cc18669d34b0da23f0d68d137","last_reissued_at":"2026-06-04T01:08:50.037051Z","signature_status":"signed_v1","first_computed_at":"2026-06-04T01:08:50.037051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Contextual Multi-Task Reinforcement Learning for Autonomous Reef Monitoring","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Contextual multi-task reinforcement learning trains one policy to handle multiple reef monitoring tasks.","cross_cats":["cs.AI"],"primary_cat":"cs.RO","authors_text":"Frank Kirchner, Mariela De Lucas Alvarez, Melvin Laux, Rebecca Adam, Rina Alo, S\\\"oren T\\\"opper, Yi-Ling Liu","submitted_at":"2026-04-14T12:16:56Z","abstract_excerpt":"Although autonomous underwater vehicles promise the capability of marine ecosystem monitoring, their deployment is fundamentally limited by the difficulty of controlling vehicles under highly uncertain and non-stationary underwater dynamics. To address these challenges, we employ a data-driven reinforcement learning approach to compensate for unknown dynamics and task variations. Traditional single-task reinforcement learning has a tendency to overfit the training environment, thus, limit the long-term usefulness of the learnt policy. Hence, we propose to use a contextual multi-task reinforcem"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We train a single context-dependent policy that is able to solve multiple related monitoring tasks in a simulated reef environment in HoloOcean.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That performance gains observed in the HoloOcean simulator will translate to real-world underwater environments with highly uncertain and non-stationary dynamics.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A context-dependent multi-task RL policy is trained and evaluated in HoloOcean simulation to solve multiple reef monitoring tasks with claimed improvements in sample efficiency, zero-shot generalization, and robustness to water currents.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Contextual multi-task reinforcement learning trains one policy to handle multiple reef monitoring tasks.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"d0fcdf051c9c3faa3c6730792595ead42e55e923da379ff3e2d053957aedbf4f"},"source":{"id":"2604.12645","kind":"arxiv","version":2},"verdict":{"id":"7cc8bb5f-551b-4202-a9da-37d9946a83cb","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-10T15:08:54.040768Z","strongest_claim":"We train a single context-dependent policy that is able to solve multiple related monitoring tasks in a simulated reef environment in HoloOcean.","one_line_summary":"A context-dependent multi-task RL policy is trained and evaluated in HoloOcean simulation to solve multiple reef monitoring tasks with claimed improvements in sample efficiency, zero-shot generalization, and robustness to water currents.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That performance gains observed in the HoloOcean simulator will translate to real-world underwater environments with highly uncertain and non-stationary dynamics.","pith_extraction_headline":"Contextual multi-task reinforcement learning trains one policy to handle multiple reef monitoring tasks."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.12645/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2604.12645","created_at":"2026-06-04T01:08:50.037153+00:00"},{"alias_kind":"arxiv_version","alias_value":"2604.12645v2","created_at":"2026-06-04T01:08:50.037153+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.12645","created_at":"2026-06-04T01:08:50.037153+00:00"},{"alias_kind":"pith_short_12","alias_value":"N3K3BDIEGEX2","created_at":"2026-06-04T01:08:50.037153+00:00"},{"alias_kind":"pith_short_16","alias_value":"N3K3BDIEGEX2C7MF","created_at":"2026-06-04T01:08:50.037153+00:00"},{"alias_kind":"pith_short_8","alias_value":"N3K3BDIE","created_at":"2026-06-04T01:08:50.037153+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2604.21640","citing_title":"Task-specific Subnetwork Discovery in Reinforcement Learning for Autonomous Underwater Navigation","ref_index":17,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN","json":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN.json","graph_json":"https://pith.science/api/pith-number/N3K3BDIEGEX2C7MFBOZL3G2WHN/graph.json","events_json":"https://pith.science/api/pith-number/N3K3BDIEGEX2C7MFBOZL3G2WHN/events.json","paper":"https://pith.science/paper/N3K3BDIE"},"agent_actions":{"view_html":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN","download_json":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN.json","view_paper":"https://pith.science/paper/N3K3BDIE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2604.12645&json=true","fetch_graph":"https://pith.science/api/pith-number/N3K3BDIEGEX2C7MFBOZL3G2WHN/graph.json","fetch_events":"https://pith.science/api/pith-number/N3K3BDIEGEX2C7MFBOZL3G2WHN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN/action/storage_attestation","attest_author":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN/action/author_attestation","sign_citation":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN/action/citation_signature","submit_replication":"https://pith.science/pith/N3K3BDIEGEX2C7MFBOZL3G2WHN/action/replication_record"}},"created_at":"2026-06-04T01:08:50.037153+00:00","updated_at":"2026-06-04T01:08:50.037153+00:00"}