{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:JKMVSCVZ2D4SPBG4IZ7DROZV6U","short_pith_number":"pith:JKMVSCVZ","canonical_record":{"source":{"id":"2504.17780","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-24T17:56:22Z","cross_cats_sorted":[],"title_canon_sha256":"6853e4132906fd4bff32cac78e51031ac925ee365bfe81a584837607e0dd6566","abstract_canon_sha256":"db4242bae9285c2ff02721d397948d44e6ef2770e45f10a5d92c246e6de77126"},"schema_version":"1.0"},"canonical_sha256":"4a99590ab9d0f92784dc467e38bb35f5375a0183e603cf475210a06fc527dc9c","source":{"kind":"arxiv","id":"2504.17780","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.17780","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"arxiv_version","alias_value":"2504.17780v1","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.17780","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"pith_short_12","alias_value":"JKMVSCVZ2D4S","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"pith_short_16","alias_value":"JKMVSCVZ2D4SPBG4","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"pith_short_8","alias_value":"JKMVSCVZ","created_at":"2026-07-05T10:53:40Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:JKMVSCVZ2D4SPBG4IZ7DROZV6U","target":"record","payload":{"canonical_record":{"source":{"id":"2504.17780","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-24T17:56:22Z","cross_cats_sorted":[],"title_canon_sha256":"6853e4132906fd4bff32cac78e51031ac925ee365bfe81a584837607e0dd6566","abstract_canon_sha256":"db4242bae9285c2ff02721d397948d44e6ef2770e45f10a5d92c246e6de77126"},"schema_version":"1.0"},"canonical_sha256":"4a99590ab9d0f92784dc467e38bb35f5375a0183e603cf475210a06fc527dc9c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:53:40.751916Z","signature_b64":"dvtm6DlDUwqho6x1qgLbjnuvu/Yumbmhiq7iYg7mBFEtfPZ1arNaQrTb5f1WZShcauUEfcov/Ns6SPe2zStuBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4a99590ab9d0f92784dc467e38bb35f5375a0183e603cf475210a06fc527dc9c","last_reissued_at":"2026-07-05T10:53:40.751445Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:53:40.751445Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2504.17780","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":"2026-07-05T10:53:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NgTkDomT4NkqVib9qEqV9sylYmZgDwFkpDGqliOKH2ggAsy6bmrBU5KkV91OWT4DN6QTfMLk7tKOgpUs5hSIBQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T18:51:01.754946Z"},"content_sha256":"9e0988a8356a74769799a28a724c9eb9306065e09c8940188230c781067226ef","schema_version":"1.0","event_id":"sha256:9e0988a8356a74769799a28a724c9eb9306065e09c8940188230c781067226ef"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:JKMVSCVZ2D4SPBG4IZ7DROZV6U","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Replay to Remember: Retaining Domain Knowledge in Streaming Language Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Sneh Pillai (University of Massachusetts Dartmouth)","submitted_at":"2025-04-24T17:56:22Z","abstract_excerpt":"Continual learning in large language models (LLMs) typically encounters the critical challenge of catastrophic forgetting, where previously acquired knowledge deteriorates upon exposure to new data. While techniques like replay buffers and parameter-efficient tuning (e.g., Low-Rank Adaptation or LoRA) have been proposed, few studies investigate real-time domain adaptation under strict computational and data-stream constraints. In this paper, we demonstrate a lightweight method combining LoRA and a minimal replay mechanism in a realistic streaming setting across three diverse knowledge domains:"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.17780","kind":"arxiv","version":1},"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/2504.17780/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T10:53:40Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"bi/evU0yuTxPiMpwesqTjHzahJh5/2qrbcH2tDgNkVQiIm5f+N8PZ41Jg/b1jCrCbOizcNl5UeHxMd/609l1Aw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-18T18:51:01.755947Z"},"content_sha256":"35ac29b4603d088d988dd2f152e7bc944c0009e3a846173c332c841535d81119","schema_version":"1.0","event_id":"sha256:35ac29b4603d088d988dd2f152e7bc944c0009e3a846173c332c841535d81119"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JKMVSCVZ2D4SPBG4IZ7DROZV6U/bundle.json","state_url":"https://pith.science/pith/JKMVSCVZ2D4SPBG4IZ7DROZV6U/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JKMVSCVZ2D4SPBG4IZ7DROZV6U/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":"2026-08-18T18:51:01Z","links":{"resolver":"https://pith.science/pith/JKMVSCVZ2D4SPBG4IZ7DROZV6U","bundle":"https://pith.science/pith/JKMVSCVZ2D4SPBG4IZ7DROZV6U/bundle.json","state":"https://pith.science/pith/JKMVSCVZ2D4SPBG4IZ7DROZV6U/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JKMVSCVZ2D4SPBG4IZ7DROZV6U/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:JKMVSCVZ2D4SPBG4IZ7DROZV6U","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":"db4242bae9285c2ff02721d397948d44e6ef2770e45f10a5d92c246e6de77126","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-24T17:56:22Z","title_canon_sha256":"6853e4132906fd4bff32cac78e51031ac925ee365bfe81a584837607e0dd6566"},"schema_version":"1.0","source":{"id":"2504.17780","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2504.17780","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"arxiv_version","alias_value":"2504.17780v1","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.17780","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"pith_short_12","alias_value":"JKMVSCVZ2D4S","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"pith_short_16","alias_value":"JKMVSCVZ2D4SPBG4","created_at":"2026-07-05T10:53:40Z"},{"alias_kind":"pith_short_8","alias_value":"JKMVSCVZ","created_at":"2026-07-05T10:53:40Z"}],"graph_snapshots":[{"event_id":"sha256:35ac29b4603d088d988dd2f152e7bc944c0009e3a846173c332c841535d81119","target":"graph","created_at":"2026-07-05T10:53:40Z","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":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2504.17780/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Continual learning in large language models (LLMs) typically encounters the critical challenge of catastrophic forgetting, where previously acquired knowledge deteriorates upon exposure to new data. While techniques like replay buffers and parameter-efficient tuning (e.g., Low-Rank Adaptation or LoRA) have been proposed, few studies investigate real-time domain adaptation under strict computational and data-stream constraints. In this paper, we demonstrate a lightweight method combining LoRA and a minimal replay mechanism in a realistic streaming setting across three diverse knowledge domains:","authors_text":"Sneh Pillai (University of Massachusetts Dartmouth)","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-24T17:56:22Z","title":"Replay to Remember: Retaining Domain Knowledge in Streaming Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.17780","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:9e0988a8356a74769799a28a724c9eb9306065e09c8940188230c781067226ef","target":"record","created_at":"2026-07-05T10:53:40Z","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":"db4242bae9285c2ff02721d397948d44e6ef2770e45f10a5d92c246e6de77126","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-04-24T17:56:22Z","title_canon_sha256":"6853e4132906fd4bff32cac78e51031ac925ee365bfe81a584837607e0dd6566"},"schema_version":"1.0","source":{"id":"2504.17780","kind":"arxiv","version":1}},"canonical_sha256":"4a99590ab9d0f92784dc467e38bb35f5375a0183e603cf475210a06fc527dc9c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4a99590ab9d0f92784dc467e38bb35f5375a0183e603cf475210a06fc527dc9c","first_computed_at":"2026-07-05T10:53:40.751445Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T10:53:40.751445Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"dvtm6DlDUwqho6x1qgLbjnuvu/Yumbmhiq7iYg7mBFEtfPZ1arNaQrTb5f1WZShcauUEfcov/Ns6SPe2zStuBw==","signature_status":"signed_v1","signed_at":"2026-07-05T10:53:40.751916Z","signed_message":"canonical_sha256_bytes"},"source_id":"2504.17780","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9e0988a8356a74769799a28a724c9eb9306065e09c8940188230c781067226ef","sha256:35ac29b4603d088d988dd2f152e7bc944c0009e3a846173c332c841535d81119"],"state_sha256":"b571047d721be49934fdb2f4d2f76bd846d7f3fa22a4b3f0db8e663638af9ab9"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"sT4wRWejQFEX1/iF9PNsHuRs2Z5gq8OUIERGJufopGgq1pnoCBzTsacmhYL5DiGYpkZfAq9W7hMP981opFSeCw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-18T18:51:01.763487Z","bundle_sha256":"454eec465b20b10b312b8507971c3bffd874bc2301d4639913774e254dc8ee9b"}}