{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:UOGG4A7BLI5EOR2T7KKL4FOXTC","short_pith_number":"pith:UOGG4A7B","canonical_record":{"source":{"id":"2604.26836","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-29T16:01:59Z","cross_cats_sorted":["cs.SY","eess.SY"],"title_canon_sha256":"0ad2ee7877817b5067a1aa9cf10890d83ef3aa155eb6f0a1ae879792379e009f","abstract_canon_sha256":"20a65ff2808ec68cfdf146124da026fc790d885c7ad080d802ffb50160174bbe"},"schema_version":"1.0"},"canonical_sha256":"a38c6e03e15a3a474753fa94be15d798a506302743d3202bf64772510573b1b1","source":{"kind":"arxiv","id":"2604.26836","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.26836","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"arxiv_version","alias_value":"2604.26836v2","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.26836","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"pith_short_12","alias_value":"UOGG4A7BLI5E","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"pith_short_16","alias_value":"UOGG4A7BLI5EOR2T","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"pith_short_8","alias_value":"UOGG4A7B","created_at":"2026-05-22T02:04:41Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:UOGG4A7BLI5EOR2T7KKL4FOXTC","target":"record","payload":{"canonical_record":{"source":{"id":"2604.26836","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-29T16:01:59Z","cross_cats_sorted":["cs.SY","eess.SY"],"title_canon_sha256":"0ad2ee7877817b5067a1aa9cf10890d83ef3aa155eb6f0a1ae879792379e009f","abstract_canon_sha256":"20a65ff2808ec68cfdf146124da026fc790d885c7ad080d802ffb50160174bbe"},"schema_version":"1.0"},"canonical_sha256":"a38c6e03e15a3a474753fa94be15d798a506302743d3202bf64772510573b1b1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-22T02:04:41.423053Z","signature_b64":"jNXf/OyFzd9erUe0zObTKp5ci/JjYlmW+Mxv8wOnkeczlcLOgdbIxyfdxNu9k49k/B4O0x14BF6BKieyh/uUAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a38c6e03e15a3a474753fa94be15d798a506302743d3202bf64772510573b1b1","last_reissued_at":"2026-05-22T02:04:41.422199Z","signature_status":"signed_v1","first_computed_at":"2026-05-22T02:04:41.422199Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2604.26836","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-22T02:04:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"dWFQd8thkR9MtlQkzeedz6COKWsKGJtBqI6ER1pkVoW0BW8f6WPTW0MwCLXkY8pEin3glgKf0anIy//uF/rNBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T09:23:56.803346Z"},"content_sha256":"881179a1eef02587e68018d66c01aca8025656083952c377c854475a035739b4","schema_version":"1.0","event_id":"sha256:881179a1eef02587e68018d66c01aca8025656083952c377c854475a035739b4"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:UOGG4A7BLI5EOR2T7KKL4FOXTC","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Uncertainty-Aware Predictive Safety Filters for Probabilistic Neural Network Dynamics","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"The Uncertainty-Aware Predictive Safety Filter uses reachable sets from probabilistic ensemble neural networks and an explicit certainty constraint to guarantee safety during model-based reinforcement learning exploration.","cross_cats":["cs.SY","eess.SY"],"primary_cat":"cs.LG","authors_text":"Bernd Frauenknecht, Daniel Mayfrank, Henrik Hose, Lukas Kesper, Sebastian Trimpe","submitted_at":"2026-04-29T16:01:59Z","abstract_excerpt":"Predictive safety filters (PSFs) leverage model predictive control to enforce constraint satisfaction during deep reinforcement learning (RL) exploration, yet their reliance on first-principles models or Gaussian processes limits scalability and broader applicability. Meanwhile, model-based RL (MBRL) methods routinely employ probabilistic ensemble (PE) neural networks to capture complex, high-dimensional dynamics from data with minimal prior knowledge. However, existing attempts to integrate PEs into PSFs lack rigorous uncertainty quantification. We introduce the Uncertainty-Aware Predictive S"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"We introduce the Uncertainty-Aware Predictive Safety Filter (UPSi), a PSF that provides rigorous safety predictions using PE dynamics models by formulating future outcomes as reachable sets. UPSi introduces an explicit certainty constraint that prevents model exploitation and integrates seamlessly into common MBRL frameworks.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That reachable sets derived from probabilistic ensemble neural network predictions can be computed rigorously enough to guarantee constraint satisfaction, and that the certainty constraint sufficiently prevents exploitation of model uncertainty without overly restricting exploration.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"UPSi integrates probabilistic neural network dynamics into predictive safety filters via reachable sets and a certainty constraint, improving safety in model-based RL exploration while matching standard performance.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"The Uncertainty-Aware Predictive Safety Filter uses reachable sets from probabilistic ensemble neural networks and an explicit certainty constraint to guarantee safety during model-based reinforcement learning exploration.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"83d9ebc9ddb1f8fd36d278e30e8dd715e75011742f36a7928969f37883d68f56"},"source":{"id":"2604.26836","kind":"arxiv","version":2},"verdict":{"id":"494e8202-a51c-4435-a351-4c8223d047f7","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-07T13:16:32.007677Z","strongest_claim":"We introduce the Uncertainty-Aware Predictive Safety Filter (UPSi), a PSF that provides rigorous safety predictions using PE dynamics models by formulating future outcomes as reachable sets. UPSi introduces an explicit certainty constraint that prevents model exploitation and integrates seamlessly into common MBRL frameworks.","one_line_summary":"UPSi integrates probabilistic neural network dynamics into predictive safety filters via reachable sets and a certainty constraint, improving safety in model-based RL exploration while matching standard performance.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That reachable sets derived from probabilistic ensemble neural network predictions can be computed rigorously enough to guarantee constraint satisfaction, and that the certainty constraint sufficiently prevents exploitation of model uncertainty without overly restricting exploration.","pith_extraction_headline":"The Uncertainty-Aware Predictive Safety Filter uses reachable sets from probabilistic ensemble neural networks and an explicit certainty constraint to guarantee safety during model-based reinforcement learning exploration."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.26836/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T23:40:39.277962Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T19:45:01.148181Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"b3b13a0800abf90908ad09d3a84de059c5aadc3e3b4b94ab310e436d76bf6442"},"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"},"verdict_id":"494e8202-a51c-4435-a351-4c8223d047f7"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-22T02:04:41Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"sf1WAalnKfH2LcTwwE2m2+gIcJJbPhAg6OYvwR0yO4PqcVE06Zw/tZIK5I9ep2CPRmLEtXvh1MqZnY5UsDMHAw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-28T09:23:56.803887Z"},"content_sha256":"0faf6bbb29ea921b69dd0fd883400d1cf05507b1301cf295370d4dd577fc4844","schema_version":"1.0","event_id":"sha256:0faf6bbb29ea921b69dd0fd883400d1cf05507b1301cf295370d4dd577fc4844"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/UOGG4A7BLI5EOR2T7KKL4FOXTC/bundle.json","state_url":"https://pith.science/pith/UOGG4A7BLI5EOR2T7KKL4FOXTC/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/UOGG4A7BLI5EOR2T7KKL4FOXTC/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":"2026-05-28T09:23:56Z","links":{"resolver":"https://pith.science/pith/UOGG4A7BLI5EOR2T7KKL4FOXTC","bundle":"https://pith.science/pith/UOGG4A7BLI5EOR2T7KKL4FOXTC/bundle.json","state":"https://pith.science/pith/UOGG4A7BLI5EOR2T7KKL4FOXTC/state.json","well_known_bundle":"https://pith.science/.well-known/pith/UOGG4A7BLI5EOR2T7KKL4FOXTC/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:UOGG4A7BLI5EOR2T7KKL4FOXTC","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":"20a65ff2808ec68cfdf146124da026fc790d885c7ad080d802ffb50160174bbe","cross_cats_sorted":["cs.SY","eess.SY"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-29T16:01:59Z","title_canon_sha256":"0ad2ee7877817b5067a1aa9cf10890d83ef3aa155eb6f0a1ae879792379e009f"},"schema_version":"1.0","source":{"id":"2604.26836","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2604.26836","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"arxiv_version","alias_value":"2604.26836v2","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2604.26836","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"pith_short_12","alias_value":"UOGG4A7BLI5E","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"pith_short_16","alias_value":"UOGG4A7BLI5EOR2T","created_at":"2026-05-22T02:04:41Z"},{"alias_kind":"pith_short_8","alias_value":"UOGG4A7B","created_at":"2026-05-22T02:04:41Z"}],"graph_snapshots":[{"event_id":"sha256:0faf6bbb29ea921b69dd0fd883400d1cf05507b1301cf295370d4dd577fc4844","target":"graph","created_at":"2026-05-22T02:04:41Z","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 introduce the Uncertainty-Aware Predictive Safety Filter (UPSi), a PSF that provides rigorous safety predictions using PE dynamics models by formulating future outcomes as reachable sets. UPSi introduces an explicit certainty constraint that prevents model exploitation and integrates seamlessly into common MBRL frameworks."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That reachable sets derived from probabilistic ensemble neural network predictions can be computed rigorously enough to guarantee constraint satisfaction, and that the certainty constraint sufficiently prevents exploitation of model uncertainty without overly restricting exploration."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"UPSi integrates probabilistic neural network dynamics into predictive safety filters via reachable sets and a certainty constraint, improving safety in model-based RL exploration while matching standard performance."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"The Uncertainty-Aware Predictive Safety Filter uses reachable sets from probabilistic ensemble neural networks and an explicit certainty constraint to guarantee safety during model-based reinforcement learning exploration."}],"snapshot_sha256":"83d9ebc9ddb1f8fd36d278e30e8dd715e75011742f36a7928969f37883d68f56"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[{"findings_count":0,"name":"ai_meta_artifact","ran_at":"2026-05-20T23:40:39.277962Z","status":"completed","version":"1.0.0"},{"findings_count":0,"name":"doi_compliance","ran_at":"2026-05-19T19:45:01.148181Z","status":"completed","version":"1.0.0"}],"endpoint":"/pith/2604.26836/integrity.json","findings":[],"snapshot_sha256":"b3b13a0800abf90908ad09d3a84de059c5aadc3e3b4b94ab310e436d76bf6442","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Predictive safety filters (PSFs) leverage model predictive control to enforce constraint satisfaction during deep reinforcement learning (RL) exploration, yet their reliance on first-principles models or Gaussian processes limits scalability and broader applicability. Meanwhile, model-based RL (MBRL) methods routinely employ probabilistic ensemble (PE) neural networks to capture complex, high-dimensional dynamics from data with minimal prior knowledge. However, existing attempts to integrate PEs into PSFs lack rigorous uncertainty quantification. We introduce the Uncertainty-Aware Predictive S","authors_text":"Bernd Frauenknecht, Daniel Mayfrank, Henrik Hose, Lukas Kesper, Sebastian Trimpe","cross_cats":["cs.SY","eess.SY"],"headline":"The Uncertainty-Aware Predictive Safety Filter uses reachable sets from probabilistic ensemble neural networks and an explicit certainty constraint to guarantee safety during model-based reinforcement learning exploration.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-29T16:01:59Z","title":"Uncertainty-Aware Predictive Safety Filters for Probabilistic Neural Network Dynamics"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2604.26836","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-07T13:16:32.007677Z","id":"494e8202-a51c-4435-a351-4c8223d047f7","model_set":{"reader":"grok-4.3"},"one_line_summary":"UPSi integrates probabilistic neural network dynamics into predictive safety filters via reachable sets and a certainty constraint, improving safety in model-based RL exploration while matching standard performance.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"The Uncertainty-Aware Predictive Safety Filter uses reachable sets from probabilistic ensemble neural networks and an explicit certainty constraint to guarantee safety during model-based reinforcement learning exploration.","strongest_claim":"We introduce the Uncertainty-Aware Predictive Safety Filter (UPSi), a PSF that provides rigorous safety predictions using PE dynamics models by formulating future outcomes as reachable sets. UPSi introduces an explicit certainty constraint that prevents model exploitation and integrates seamlessly into common MBRL frameworks.","weakest_assumption":"That reachable sets derived from probabilistic ensemble neural network predictions can be computed rigorously enough to guarantee constraint satisfaction, and that the certainty constraint sufficiently prevents exploitation of model uncertainty without overly restricting exploration."}},"verdict_id":"494e8202-a51c-4435-a351-4c8223d047f7"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:881179a1eef02587e68018d66c01aca8025656083952c377c854475a035739b4","target":"record","created_at":"2026-05-22T02:04:41Z","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":"20a65ff2808ec68cfdf146124da026fc790d885c7ad080d802ffb50160174bbe","cross_cats_sorted":["cs.SY","eess.SY"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-04-29T16:01:59Z","title_canon_sha256":"0ad2ee7877817b5067a1aa9cf10890d83ef3aa155eb6f0a1ae879792379e009f"},"schema_version":"1.0","source":{"id":"2604.26836","kind":"arxiv","version":2}},"canonical_sha256":"a38c6e03e15a3a474753fa94be15d798a506302743d3202bf64772510573b1b1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"a38c6e03e15a3a474753fa94be15d798a506302743d3202bf64772510573b1b1","first_computed_at":"2026-05-22T02:04:41.422199Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-22T02:04:41.422199Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"jNXf/OyFzd9erUe0zObTKp5ci/JjYlmW+Mxv8wOnkeczlcLOgdbIxyfdxNu9k49k/B4O0x14BF6BKieyh/uUAQ==","signature_status":"signed_v1","signed_at":"2026-05-22T02:04:41.423053Z","signed_message":"canonical_sha256_bytes"},"source_id":"2604.26836","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:881179a1eef02587e68018d66c01aca8025656083952c377c854475a035739b4","sha256:0faf6bbb29ea921b69dd0fd883400d1cf05507b1301cf295370d4dd577fc4844"],"state_sha256":"e06080a9e480a0f3b499b9baf4a3b93b094867c1dfe21c610c5540c66d27615f"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"A8FOxMrwqquIJQaUr7bOe3v/Z4rds3QqDsHm0g9e6derKLeQd5aaA/lBi3XigdMOTQV9D0A3RwgcI3nmmBwkCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-28T09:23:56.806460Z","bundle_sha256":"2113259a70b3f4ec84c0cfdd6af95a2eb5f3095cd5170aa1c3ef307d59663c3c"}}