{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2026:JNGWELOJJIN54L4MFZ5ZKJQGK7","short_pith_number":"pith:JNGWELOJ","canonical_record":{"source":{"id":"2605.13678","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T15:36:46Z","cross_cats_sorted":[],"title_canon_sha256":"b6447f6619c3d14f3271e2e41ef50acb138f8a4b444eda86ee18f1670118bb2b","abstract_canon_sha256":"45153ef9997ab91657af8cc565a0bfd0cebb8e558f8c15fa325686839bda774e"},"schema_version":"1.0"},"canonical_sha256":"4b4d622dc94a1bde2f8c2e7b95260657d7e5973ce942a87ed73e5f0c42ab8b1b","source":{"kind":"arxiv","id":"2605.13678","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13678","created_at":"2026-05-18T02:44:17Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13678v1","created_at":"2026-05-18T02:44:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13678","created_at":"2026-05-18T02:44:17Z"},{"alias_kind":"pith_short_12","alias_value":"JNGWELOJJIN5","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"JNGWELOJJIN54L4M","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"JNGWELOJ","created_at":"2026-05-18T12:33:37Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2026:JNGWELOJJIN54L4MFZ5ZKJQGK7","target":"record","payload":{"canonical_record":{"source":{"id":"2605.13678","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T15:36:46Z","cross_cats_sorted":[],"title_canon_sha256":"b6447f6619c3d14f3271e2e41ef50acb138f8a4b444eda86ee18f1670118bb2b","abstract_canon_sha256":"45153ef9997ab91657af8cc565a0bfd0cebb8e558f8c15fa325686839bda774e"},"schema_version":"1.0"},"canonical_sha256":"4b4d622dc94a1bde2f8c2e7b95260657d7e5973ce942a87ed73e5f0c42ab8b1b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T02:44:17.056799Z","signature_b64":"8/sqRBheTfL1INS1/InVOvDVTItUoqXXpFtdzROzrXe9Gqv0lCpX4rd9tPwSzowPhjiNspnklBvqGtFW/+AuAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4b4d622dc94a1bde2f8c2e7b95260657d7e5973ce942a87ed73e5f0c42ab8b1b","last_reissued_at":"2026-05-18T02:44:17.056262Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T02:44:17.056262Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2605.13678","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":"2026-05-18T02:44:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+pS4VXj25XhoLEsfEpLIWxZyKJbHyU/lTJ2FdP+FKT4n+hSNjashcdDT+88T7mr9WhRrFFubnSIpEt/ihXf3Bw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-05T02:54:43.634019Z"},"content_sha256":"c8abe6d312dc1a9056e3ea362186cc6319bca0ea8cdc865ef23afa505cd6d566","schema_version":"1.0","event_id":"sha256:c8abe6d312dc1a9056e3ea362186cc6319bca0ea8cdc865ef23afa505cd6d566"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2026:JNGWELOJJIN54L4MFZ5ZKJQGK7","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Three-Stage Learning Unlocks Strong Performance in Simple Models for Long-Term Time Series Forecasting","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A three-stage training process lets simple MLP models match complex architectures on long-term time series forecasting.","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Guangxin Jiang, Jin Yang, Zhenan Yu","submitted_at":"2026-05-13T15:36:46Z","abstract_excerpt":"Recent studies on long-term time series forecasting have shown that simple linear models and MLP-based predictors can achieve strong performance without increasingly complex architectures. However, many competitive baselines still rely on structural priors such as frequency-domain modeling, explicit decomposition, multi-scale mixing, or sophisticated cross-variable interaction modules, while paying less attention to how simple temporal mappings should be trained and organized. In this paper, we propose STAIR, short for Stagewise Temporal Adaptation via Individualization and Residual Learning, "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Experiments on nine long-term forecasting benchmarks show that STAIR matches or outperforms recent strong baselines while preserving a simple temporal backbone.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the three stages can be trained sequentially on a shallow MLP without later stages overwriting useful shared patterns learned earlier, and that the simple backbone has enough capacity to benefit from the added individualization and residual components.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"STAIR's three-stage training enables simple temporal models to match or exceed complex baselines on long-term forecasting benchmarks by combining shared learning, individual adaptation, and residual cross-variable modeling.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A three-stage training process lets simple MLP models match complex architectures on long-term time series forecasting.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"19a933490ea6e54d8a7aa8fbb146a4d98471de9860f2e1ae6e31fe051fd6307b"},"source":{"id":"2605.13678","kind":"arxiv","version":1},"verdict":{"id":"e83446ea-dc22-468a-94fb-f3fb7fc4cd29","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T19:08:45.775161Z","strongest_claim":"Experiments on nine long-term forecasting benchmarks show that STAIR matches or outperforms recent strong baselines while preserving a simple temporal backbone.","one_line_summary":"STAIR's three-stage training enables simple temporal models to match or exceed complex baselines on long-term forecasting benchmarks by combining shared learning, individual adaptation, and residual cross-variable modeling.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the three stages can be trained sequentially on a shallow MLP without later stages overwriting useful shared patterns learned earlier, and that the simple backbone has enough capacity to benefit from the added individualization and residual components.","pith_extraction_headline":"A three-stage training process lets simple MLP models match complex architectures on long-term time series forecasting."},"references":{"count":17,"sample":[{"doi":"","year":null,"title":"Proceedings of the AAAI Conference on Artificial Intelligence , year=","work_id":"ee0d4b2a-1b5a-4380-be86-4a5bb92bb8cc","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Advances in Neural Information Processing Systems , year=","work_id":"4b2e7438-9d39-4d61-aa5b-fb354311af60","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"International Conference on Machine Learning , year=","work_id":"90ce4a3b-8536-48b9-ac84-80c5ee03e8b0","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"International Conference on Learning Representations , year=","work_id":"acb99310-e811-41ab-a0d6-768bda1a7d4b","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Proceedings of the AAAI Conference on Artificial Intelligence , year=","work_id":"ebfd9442-627c-4ffc-8f3f-3e1c1276a51d","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":17,"snapshot_sha256":"807b7bb434e216b34132c647aa7792c6fa839f57665beac7c9783a42d1f4bde1","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":"e83446ea-dc22-468a-94fb-f3fb7fc4cd29"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T02:44:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"4GQhXQqwsQXQNOIUm31UK4Www4PLirydWMBlD4ELxspkY4FTr/AwBRshwY/Y3XR979CQhQslg8gpDwjPeUrjBA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-05T02:54:43.634996Z"},"content_sha256":"c817312692e06c4d1abbfaf9d7eae1f7a37b6bb7195a9874fa417dfffea7a378","schema_version":"1.0","event_id":"sha256:c817312692e06c4d1abbfaf9d7eae1f7a37b6bb7195a9874fa417dfffea7a378"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/JNGWELOJJIN54L4MFZ5ZKJQGK7/bundle.json","state_url":"https://pith.science/pith/JNGWELOJJIN54L4MFZ5ZKJQGK7/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/JNGWELOJJIN54L4MFZ5ZKJQGK7/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":"2026-06-05T02:54:43Z","links":{"resolver":"https://pith.science/pith/JNGWELOJJIN54L4MFZ5ZKJQGK7","bundle":"https://pith.science/pith/JNGWELOJJIN54L4MFZ5ZKJQGK7/bundle.json","state":"https://pith.science/pith/JNGWELOJJIN54L4MFZ5ZKJQGK7/state.json","well_known_bundle":"https://pith.science/.well-known/pith/JNGWELOJJIN54L4MFZ5ZKJQGK7/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:JNGWELOJJIN54L4MFZ5ZKJQGK7","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":"45153ef9997ab91657af8cc565a0bfd0cebb8e558f8c15fa325686839bda774e","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T15:36:46Z","title_canon_sha256":"b6447f6619c3d14f3271e2e41ef50acb138f8a4b444eda86ee18f1670118bb2b"},"schema_version":"1.0","source":{"id":"2605.13678","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13678","created_at":"2026-05-18T02:44:17Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13678v1","created_at":"2026-05-18T02:44:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13678","created_at":"2026-05-18T02:44:17Z"},{"alias_kind":"pith_short_12","alias_value":"JNGWELOJJIN5","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"JNGWELOJJIN54L4M","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"JNGWELOJ","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:c817312692e06c4d1abbfaf9d7eae1f7a37b6bb7195a9874fa417dfffea7a378","target":"graph","created_at":"2026-05-18T02:44:17Z","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":"Experiments on nine long-term forecasting benchmarks show that STAIR matches or outperforms recent strong baselines while preserving a simple temporal backbone."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"That the three stages can be trained sequentially on a shallow MLP without later stages overwriting useful shared patterns learned earlier, and that the simple backbone has enough capacity to benefit from the added individualization and residual components."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"STAIR's three-stage training enables simple temporal models to match or exceed complex baselines on long-term forecasting benchmarks by combining shared learning, individual adaptation, and residual cross-variable modeling."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A three-stage training process lets simple MLP models match complex architectures on long-term time series forecasting."}],"snapshot_sha256":"19a933490ea6e54d8a7aa8fbb146a4d98471de9860f2e1ae6e31fe051fd6307b"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"paper":{"abstract_excerpt":"Recent studies on long-term time series forecasting have shown that simple linear models and MLP-based predictors can achieve strong performance without increasingly complex architectures. However, many competitive baselines still rely on structural priors such as frequency-domain modeling, explicit decomposition, multi-scale mixing, or sophisticated cross-variable interaction modules, while paying less attention to how simple temporal mappings should be trained and organized. In this paper, we propose STAIR, short for Stagewise Temporal Adaptation via Individualization and Residual Learning, ","authors_text":"Guangxin Jiang, Jin Yang, Zhenan Yu","cross_cats":[],"headline":"A three-stage training process lets simple MLP models match complex architectures on long-term time series forecasting.","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T15:36:46Z","title":"Three-Stage Learning Unlocks Strong Performance in Simple Models for Long-Term Time Series Forecasting"},"references":{"count":17,"internal_anchors":0,"resolved_work":17,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"Proceedings of the AAAI Conference on Artificial Intelligence , year=","work_id":"ee0d4b2a-1b5a-4380-be86-4a5bb92bb8cc","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"Advances in Neural Information Processing Systems , year=","work_id":"4b2e7438-9d39-4d61-aa5b-fb354311af60","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"International Conference on Machine Learning , year=","work_id":"90ce4a3b-8536-48b9-ac84-80c5ee03e8b0","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"International Conference on Learning Representations , year=","work_id":"acb99310-e811-41ab-a0d6-768bda1a7d4b","year":null},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"Proceedings of the AAAI Conference on Artificial Intelligence , year=","work_id":"ebfd9442-627c-4ffc-8f3f-3e1c1276a51d","year":null}],"snapshot_sha256":"807b7bb434e216b34132c647aa7792c6fa839f57665beac7c9783a42d1f4bde1"},"source":{"id":"2605.13678","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T19:08:45.775161Z","id":"e83446ea-dc22-468a-94fb-f3fb7fc4cd29","model_set":{"reader":"grok-4.3"},"one_line_summary":"STAIR's three-stage training enables simple temporal models to match or exceed complex baselines on long-term forecasting benchmarks by combining shared learning, individual adaptation, and residual cross-variable modeling.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A three-stage training process lets simple MLP models match complex architectures on long-term time series forecasting.","strongest_claim":"Experiments on nine long-term forecasting benchmarks show that STAIR matches or outperforms recent strong baselines while preserving a simple temporal backbone.","weakest_assumption":"That the three stages can be trained sequentially on a shallow MLP without later stages overwriting useful shared patterns learned earlier, and that the simple backbone has enough capacity to benefit from the added individualization and residual components."}},"verdict_id":"e83446ea-dc22-468a-94fb-f3fb7fc4cd29"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:c8abe6d312dc1a9056e3ea362186cc6319bca0ea8cdc865ef23afa505cd6d566","target":"record","created_at":"2026-05-18T02:44:17Z","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":"45153ef9997ab91657af8cc565a0bfd0cebb8e558f8c15fa325686839bda774e","cross_cats_sorted":[],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2026-05-13T15:36:46Z","title_canon_sha256":"b6447f6619c3d14f3271e2e41ef50acb138f8a4b444eda86ee18f1670118bb2b"},"schema_version":"1.0","source":{"id":"2605.13678","kind":"arxiv","version":1}},"canonical_sha256":"4b4d622dc94a1bde2f8c2e7b95260657d7e5973ce942a87ed73e5f0c42ab8b1b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4b4d622dc94a1bde2f8c2e7b95260657d7e5973ce942a87ed73e5f0c42ab8b1b","first_computed_at":"2026-05-18T02:44:17.056262Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:17.056262Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"8/sqRBheTfL1INS1/InVOvDVTItUoqXXpFtdzROzrXe9Gqv0lCpX4rd9tPwSzowPhjiNspnklBvqGtFW/+AuAA==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:17.056799Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13678","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c8abe6d312dc1a9056e3ea362186cc6319bca0ea8cdc865ef23afa505cd6d566","sha256:c817312692e06c4d1abbfaf9d7eae1f7a37b6bb7195a9874fa417dfffea7a378"],"state_sha256":"c52572dedd5deecb3469ffd6783298807bbd6ad94825b633cce152e9a2244aa5"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"WNNpS8n3l5XiKL9EevL90kxB++SYkWLz+/69tQGOBwDQvEXIEiOBMU0HyXUkKR7ewHmD2DH3fMGjMF9WMUxkAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-05T02:54:43.638946Z","bundle_sha256":"e3c8a63d1c362d0f489cb8128bc0bd0eb4e22eafe6e3784de011e98997ffa69d"}}