{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2025:YED6Q7TAP6E7EVEYTLTFZGSAH6","short_pith_number":"pith:YED6Q7TA","canonical_record":{"source":{"id":"2501.05559","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-09T20:11:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b30b03c99602c28c93d10fcaa43418ff518ed169131b1c9ac4470e5a96fd0060","abstract_canon_sha256":"7988a4d84988c88521d0845cca33a46950ed081f8c5b02f38b7b5e3b812718ed"},"schema_version":"1.0"},"canonical_sha256":"c107e87e607f89f254989ae65c9a403f97a3143e1fb427aadc9c047f4a81b232","source":{"kind":"arxiv","id":"2501.05559","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.05559","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"arxiv_version","alias_value":"2501.05559v1","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.05559","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"pith_short_12","alias_value":"YED6Q7TAP6E7","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"pith_short_16","alias_value":"YED6Q7TAP6E7EVEY","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"pith_short_8","alias_value":"YED6Q7TA","created_at":"2026-07-05T09:59:24Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2025:YED6Q7TAP6E7EVEYTLTFZGSAH6","target":"record","payload":{"canonical_record":{"source":{"id":"2501.05559","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-09T20:11:08Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"b30b03c99602c28c93d10fcaa43418ff518ed169131b1c9ac4470e5a96fd0060","abstract_canon_sha256":"7988a4d84988c88521d0845cca33a46950ed081f8c5b02f38b7b5e3b812718ed"},"schema_version":"1.0"},"canonical_sha256":"c107e87e607f89f254989ae65c9a403f97a3143e1fb427aadc9c047f4a81b232","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:59:24.975775Z","signature_b64":"d1ZG6pJJIh0Jm7EpIOtPmWmaMB4GrXM0s6Xkzwf9U6gJbNOIefnfEiZPrhNNBpDwtNMdUTpwbv7DTbagumffCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c107e87e607f89f254989ae65c9a403f97a3143e1fb427aadc9c047f4a81b232","last_reissued_at":"2026-07-05T09:59:24.975384Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:59:24.975384Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2501.05559","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":"2026-07-05T09:59:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QH93LNSuoJLNZxbnJqjPA9ONg6MY2x7InNXojyvI+LLAMnSwenqr19Zjyal8BmcrrnKDIOYcfygvWcy64ZO7Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T09:46:24.779006Z"},"content_sha256":"2263c1d7112deb131faf9efc81c26d912632edd39b61083a2169ee735c243f33","schema_version":"1.0","event_id":"sha256:2263c1d7112deb131faf9efc81c26d912632edd39b61083a2169ee735c243f33"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2025:YED6Q7TAP6E7EVEYTLTFZGSAH6","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Soup to go: mitigating forgetting during continual learning with model averaging","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.LG","authors_text":"Anat Kleiman, Gintare Karolina Dziugaite, Jonathan Frankle, Mansheej Paul, Sham Kakade","submitted_at":"2025-01-09T20:11:08Z","abstract_excerpt":"In continual learning, where task data arrives in a sequence, fine-tuning on later tasks will often lead to performance degradation on earlier tasks. This is especially pronounced when these tasks come from diverse domains. In this setting, how can we mitigate catastrophic forgetting of earlier tasks and retain what the model has learned with minimal computational expenses? Inspired by other merging methods, and L2-regression, we propose Sequential Fine-tuning with Averaging (SFA), a method that merges currently training models with earlier checkpoints during the course of training. SOTA appro"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.05559","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/2501.05559/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":"2026-07-05T09:59:24Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"/GdPD6zcE1VI5rmF276BHPtIeFbo0QRMhK7j6IGnsjGyILygrj/f1AflVvkIEzW10qE12oqiVeguGh5y+92sDg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T09:46:24.779520Z"},"content_sha256":"48a99723631fbd767451d9c9870e224b80d7c0db6613ebae1db53bb999d5b202","schema_version":"1.0","event_id":"sha256:48a99723631fbd767451d9c9870e224b80d7c0db6613ebae1db53bb999d5b202"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/YED6Q7TAP6E7EVEYTLTFZGSAH6/bundle.json","state_url":"https://pith.science/pith/YED6Q7TAP6E7EVEYTLTFZGSAH6/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/YED6Q7TAP6E7EVEYTLTFZGSAH6/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":"2026-08-09T09:46:24Z","links":{"resolver":"https://pith.science/pith/YED6Q7TAP6E7EVEYTLTFZGSAH6","bundle":"https://pith.science/pith/YED6Q7TAP6E7EVEYTLTFZGSAH6/bundle.json","state":"https://pith.science/pith/YED6Q7TAP6E7EVEYTLTFZGSAH6/state.json","well_known_bundle":"https://pith.science/.well-known/pith/YED6Q7TAP6E7EVEYTLTFZGSAH6/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:YED6Q7TAP6E7EVEYTLTFZGSAH6","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":"7988a4d84988c88521d0845cca33a46950ed081f8c5b02f38b7b5e3b812718ed","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-09T20:11:08Z","title_canon_sha256":"b30b03c99602c28c93d10fcaa43418ff518ed169131b1c9ac4470e5a96fd0060"},"schema_version":"1.0","source":{"id":"2501.05559","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2501.05559","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"arxiv_version","alias_value":"2501.05559v1","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2501.05559","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"pith_short_12","alias_value":"YED6Q7TAP6E7","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"pith_short_16","alias_value":"YED6Q7TAP6E7EVEY","created_at":"2026-07-05T09:59:24Z"},{"alias_kind":"pith_short_8","alias_value":"YED6Q7TA","created_at":"2026-07-05T09:59:24Z"}],"graph_snapshots":[{"event_id":"sha256:48a99723631fbd767451d9c9870e224b80d7c0db6613ebae1db53bb999d5b202","target":"graph","created_at":"2026-07-05T09:59:24Z","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/2501.05559/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"In continual learning, where task data arrives in a sequence, fine-tuning on later tasks will often lead to performance degradation on earlier tasks. This is especially pronounced when these tasks come from diverse domains. In this setting, how can we mitigate catastrophic forgetting of earlier tasks and retain what the model has learned with minimal computational expenses? Inspired by other merging methods, and L2-regression, we propose Sequential Fine-tuning with Averaging (SFA), a method that merges currently training models with earlier checkpoints during the course of training. SOTA appro","authors_text":"Anat Kleiman, Gintare Karolina Dziugaite, Jonathan Frankle, Mansheej Paul, Sham Kakade","cross_cats":["cs.AI"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-09T20:11:08Z","title":"Soup to go: mitigating forgetting during continual learning with model averaging"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2501.05559","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":"sha256:2263c1d7112deb131faf9efc81c26d912632edd39b61083a2169ee735c243f33","target":"record","created_at":"2026-07-05T09:59:24Z","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":"7988a4d84988c88521d0845cca33a46950ed081f8c5b02f38b7b5e3b812718ed","cross_cats_sorted":["cs.AI"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-01-09T20:11:08Z","title_canon_sha256":"b30b03c99602c28c93d10fcaa43418ff518ed169131b1c9ac4470e5a96fd0060"},"schema_version":"1.0","source":{"id":"2501.05559","kind":"arxiv","version":1}},"canonical_sha256":"c107e87e607f89f254989ae65c9a403f97a3143e1fb427aadc9c047f4a81b232","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"c107e87e607f89f254989ae65c9a403f97a3143e1fb427aadc9c047f4a81b232","first_computed_at":"2026-07-05T09:59:24.975384Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:59:24.975384Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"d1ZG6pJJIh0Jm7EpIOtPmWmaMB4GrXM0s6Xkzwf9U6gJbNOIefnfEiZPrhNNBpDwtNMdUTpwbv7DTbagumffCw==","signature_status":"signed_v1","signed_at":"2026-07-05T09:59:24.975775Z","signed_message":"canonical_sha256_bytes"},"source_id":"2501.05559","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2263c1d7112deb131faf9efc81c26d912632edd39b61083a2169ee735c243f33","sha256:48a99723631fbd767451d9c9870e224b80d7c0db6613ebae1db53bb999d5b202"],"state_sha256":"ae3c1a5acafd2dee7427e3dfec0f1f391ccce61e27b33676f9c39fce2fe3b945"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"z2XIgmxevcxSRWWqvwhOCfATVxqbnQuD9BhjNkgrLJ7/r+B2ibwohyobKJJo5qtimg1es5sudnCeLJ7XPsXwBg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T09:46:24.785320Z","bundle_sha256":"97cb540e358d3ed0872c177d53acf06c71e3615ae3c1bfb5f717c257a26250ba"}}