{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:GVKAVV6AL3ZZQ564WHBRIQZD6J","short_pith_number":"pith:GVKAVV6A","schema_version":"1.0","canonical_sha256":"35540ad7c05ef39877dcb1c3144323f26ccc372520960a436ada129c73132780","source":{"kind":"arxiv","id":"2407.11070","version":2},"attestation_state":"computed","paper":{"title":"Optimal Defender Strategies for CAGE-2 using Causal Modeling and Tree Search","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Kim Hammar, Neil Dhir, Rolf Stadler","submitted_at":"2024-07-12T18:34:55Z","abstract_excerpt":"The CAGE-2 challenge is considered a standard benchmark to compare methods for autonomous cyber defense. Current state-of-the-art methods evaluated against this benchmark are based on model-free (offline) reinforcement learning, which does not provide provably optimal defender strategies. We address this limitation and present a formal (causal) model of CAGE-2 together with a method that produces a provably optimal defender strategy, which we call Causal Partially Observable Monte-Carlo Planning (C-POMCP). It has two key properties. First, it incorporates the causal structure of the target sys"},"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":"2407.11070","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-12T18:34:55Z","cross_cats_sorted":["cs.AI","cs.CR"],"title_canon_sha256":"6c80484799af663ca5783079dbe3404e6919598881cdb6968f9743f7e5ce7eae","abstract_canon_sha256":"586a10742b5df7a004df6a6dc549667b6661ef3f2f2c59f4625092aedbc42c56"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:23:54.588137Z","signature_b64":"qWxis3o4UMenll3gg5iBRv/EATM1uhk7rvHMlq3YBio2OHsDNXXAabGJ1+SrkEM+o4rGVedjGPqRO7Enflz7DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"35540ad7c05ef39877dcb1c3144323f26ccc372520960a436ada129c73132780","last_reissued_at":"2026-07-05T09:23:54.587664Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:23:54.587664Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimal Defender Strategies for CAGE-2 using Causal Modeling and Tree Search","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.AI","cs.CR"],"primary_cat":"cs.LG","authors_text":"Kim Hammar, Neil Dhir, Rolf Stadler","submitted_at":"2024-07-12T18:34:55Z","abstract_excerpt":"The CAGE-2 challenge is considered a standard benchmark to compare methods for autonomous cyber defense. Current state-of-the-art methods evaluated against this benchmark are based on model-free (offline) reinforcement learning, which does not provide provably optimal defender strategies. We address this limitation and present a formal (causal) model of CAGE-2 together with a method that produces a provably optimal defender strategy, which we call Causal Partially Observable Monte-Carlo Planning (C-POMCP). It has two key properties. First, it incorporates the causal structure of the target sys"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.11070","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2407.11070/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":"2407.11070","created_at":"2026-07-05T09:23:54.587722+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.11070v2","created_at":"2026-07-05T09:23:54.587722+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.11070","created_at":"2026-07-05T09:23:54.587722+00:00"},{"alias_kind":"pith_short_12","alias_value":"GVKAVV6AL3ZZ","created_at":"2026-07-05T09:23:54.587722+00:00"},{"alias_kind":"pith_short_16","alias_value":"GVKAVV6AL3ZZQ564","created_at":"2026-07-05T09:23:54.587722+00:00"},{"alias_kind":"pith_short_8","alias_value":"GVKAVV6A","created_at":"2026-07-05T09:23:54.587722+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.30479","citing_title":"COHORT: Collaborative Orchestration for Hardening via Offensive Replay on Emulated Topologies","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16205","citing_title":"Context, Reasoning, and Hierarchy: A Cost-Performance Study of Compound LLM Agent Design in an Adversarial POMDP","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2507.15163","citing_title":"Adaptive Network Security Policies via Belief Aggregation and Rollout","ref_index":43,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J","json":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J.json","graph_json":"https://pith.science/api/pith-number/GVKAVV6AL3ZZQ564WHBRIQZD6J/graph.json","events_json":"https://pith.science/api/pith-number/GVKAVV6AL3ZZQ564WHBRIQZD6J/events.json","paper":"https://pith.science/paper/GVKAVV6A"},"agent_actions":{"view_html":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J","download_json":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J.json","view_paper":"https://pith.science/paper/GVKAVV6A","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.11070&json=true","fetch_graph":"https://pith.science/api/pith-number/GVKAVV6AL3ZZQ564WHBRIQZD6J/graph.json","fetch_events":"https://pith.science/api/pith-number/GVKAVV6AL3ZZQ564WHBRIQZD6J/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J/action/storage_attestation","attest_author":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J/action/author_attestation","sign_citation":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J/action/citation_signature","submit_replication":"https://pith.science/pith/GVKAVV6AL3ZZQ564WHBRIQZD6J/action/replication_record"}},"created_at":"2026-07-05T09:23:54.587722+00:00","updated_at":"2026-07-05T09:23:54.587722+00:00"}