{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:6NIWHZOOM67FZJ3SCGCKM7LKM4","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":"c1a66aaeed12fe61c691610188d7728b89f113ee2e29c05e8f1b7dc175dde62e","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T08:58:24Z","title_canon_sha256":"8fc5c0bd06529dc31057da8642c41e4b715bcb901e7e6b93b57560bab1a2208b"},"schema_version":"1.0","source":{"id":"2605.14587","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.14587","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"arxiv_version","alias_value":"2605.14587v1","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.14587","created_at":"2026-05-17T23:39:05Z"},{"alias_kind":"pith_short_12","alias_value":"6NIWHZOOM67F","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"6NIWHZOOM67FZJ3S","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"6NIWHZOO","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:79b90572147ceb115260e210996c108c0a4a385ca8b3844ff523db17ae51a256","target":"graph","created_at":"2026-05-17T23:39:05Z","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":"only one intervention (i.e., SAM) exacerbates backdoor threats, while other interventions mitigate them. Pathological analysis identifies that the exacerbation is attributed to backdoor gradient amplification, while the mitigation stems from activation pathway disruption and representation space compression."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The 14,664 tested cases sufficiently represent the space of practical DRL deployments and attack scenarios so that the observed patterns generalize beyond the chosen environments and models."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Most plasticity interventions in DRL reduce backdoor attack success rates while SAM increases them via gradient amplification; the work introduces an SCC framework and loss-sharpness detection indicator."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Most plasticity interventions reduce backdoor threats in deep reinforcement learning, but SAM makes them worse by amplifying gradients."}],"snapshot_sha256":"f062e0a830d417c25c78fedee9ff0e787f213522cd84bf1b5891eaa46df140cc"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"50a3c357d927fa20c1923af3814d322d3d14c09fd070d118fa6cd2297ce55f66"},"paper":{"abstract_excerpt":"Extensive research has highlighted the severe threats posed by backdoor attacks to deep reinforcement learning (DRL). However, prior studies primarily focus on vanilla scenarios, while plasticity interventions have emerged as indispensable built-in components of modern DRL agents. Despite their effectiveness in mitigating plasticity loss, the impact of these interventions on DRL backdoor vulnerabilities remains underexplored, and this lack of systematic investigation poses risks in practical DRL deployments. To bridge this gap, we empirically study 14,664 cases integrating representative inter","authors_text":"Chunyi Zhou, Jiahao Chen, Linkang Du, Oubo Ma, Ruixiao Lin, Shouling Ji, Yang Dai","cross_cats":["cs.AI","cs.CR"],"headline":"Most plasticity interventions reduce backdoor threats in deep reinforcement learning, but SAM makes them worse by amplifying gradients.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T08:58:24Z","title":"Angel or Demon: Investigating the Plasticity Interventions' Impact on Backdoor Threats in Deep Reinforcement Learning"},"references":{"count":75,"internal_anchors":0,"resolved_work":75,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Continuous control with deep reinforcement learning , author=. ICLR , year=","work_id":"c29366cd-a00e-4aea-8f83-401d304795e9","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Multi-agent actor-critic for mixed cooperative-competitive environments , author=. NeurIPS , year=","work_id":"b6412379-39a2-413b-ab30-f764fa017a9e","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"CLIBE: detecting dynamic backdoors in transformer-based NLP models , author=. NDSS , year=","work_id":"c51f43cc-36ef-4301-8779-b6d58193e73d","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"IEEE transactions on neural networks and learning systems , year=","work_id":"c80309f5-3dff-4ce4-870b-aeba8bc70ed6","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Understanding black-box predictions via influence functions , author=. ICML , year=","work_id":"d9f0d1fb-12b3-4731-8c63-a7fc54900440","year":null}],"snapshot_sha256":"292cc24370876af4d00e60a0d3b9da7a6ad03033456e02c74d64185100cf7480"},"source":{"id":"2605.14587","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-15T01:33:25.686471Z","id":"a1f50100-d284-495a-ae1f-9c37a3fe9dc0","model_set":{"reader":"grok-4.3"},"one_line_summary":"Most plasticity interventions in DRL reduce backdoor attack success rates while SAM increases them via gradient amplification; the work introduces an SCC framework and loss-sharpness detection indicator.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Most plasticity interventions reduce backdoor threats in deep reinforcement learning, but SAM makes them worse by amplifying gradients.","strongest_claim":"only one intervention (i.e., SAM) exacerbates backdoor threats, while other interventions mitigate them. Pathological analysis identifies that the exacerbation is attributed to backdoor gradient amplification, while the mitigation stems from activation pathway disruption and representation space compression.","weakest_assumption":"The 14,664 tested cases sufficiently represent the space of practical DRL deployments and attack scenarios so that the observed patterns generalize beyond the chosen environments and models."}},"verdict_id":"a1f50100-d284-495a-ae1f-9c37a3fe9dc0"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:1bb10bd12362dfab088a7a58f5171796398c0160d38c8e78ce70d59f97402011","target":"record","created_at":"2026-05-17T23:39:05Z","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":"c1a66aaeed12fe61c691610188d7728b89f113ee2e29c05e8f1b7dc175dde62e","cross_cats_sorted":["cs.AI","cs.CR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T08:58:24Z","title_canon_sha256":"8fc5c0bd06529dc31057da8642c41e4b715bcb901e7e6b93b57560bab1a2208b"},"schema_version":"1.0","source":{"id":"2605.14587","kind":"arxiv","version":1}},"canonical_sha256":"f35163e5ce67be5ca7721184a67d6a67050855ee3daa3715030ad55375c26cf2","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f35163e5ce67be5ca7721184a67d6a67050855ee3daa3715030ad55375c26cf2","first_computed_at":"2026-05-17T23:39:05.291373Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-17T23:39:05.291373Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"k7LfLjwh7UQz8nVjx/jFmH6VcqiK5OTT+DEuHrmrqL/bjSAuJYNNV41OPI9DSZ2wnhJRs/ydq1zO1BQEfnpqAA==","signature_status":"signed_v1","signed_at":"2026-05-17T23:39:05.292142Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.14587","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:1bb10bd12362dfab088a7a58f5171796398c0160d38c8e78ce70d59f97402011","sha256:79b90572147ceb115260e210996c108c0a4a385ca8b3844ff523db17ae51a256"],"state_sha256":"238645ed2650fb54212f21728338b58a9678f1685d41f70ee5056fe3389397af"}