{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:WEW6ZIVGDCGHTJQE7KO5NTFMMB","short_pith_number":"pith:WEW6ZIVG","schema_version":"1.0","canonical_sha256":"b12deca2a6188c79a604fa9dd6ccac606bca2d6bbf10dd5ac6b42f2c0d9ffb4c","source":{"kind":"arxiv","id":"2503.12303","version":6},"attestation_state":"computed","paper":{"title":"Will Pre-Training Ever End? A First Step Toward Next-Generation Foundation MLLMs via Self-Improving Systematic Cognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chengyue Wu, Chi Chen, Da Peng, Helen Meng, Jen-tse Huang, Maosong Sun, Wei Ke, Xiaoying Zhang, Yipeng Zhang, Zonghao Guo","submitted_at":"2025-03-16T00:25:13Z","abstract_excerpt":"Recent progress in (multimodal) large language models ((M)LLMs) has shifted focus from pre-training to inference-time computation and post-training optimization, largely due to concerns over the availability of high-quality human data. However, these strategies alone are insufficient to drive substantial model improvements. We argue that effective model advancement requires strong synergy among pre-training, inference-time computation, and post-training optimization. In this paper, we introduce Self-Improving cognition (SIcog), a self-learning framework for constructing next-generation foundat"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2503.12303","kind":"arxiv","version":6},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-16T00:25:13Z","cross_cats_sorted":[],"title_canon_sha256":"0600d2f97b5713f921bb6a71188593b20fbfc78241498c8d702f62b145b68130","abstract_canon_sha256":"4123a5a5d24082d8686d6ea064065fc9fa33fe4550efa91a33605432dbb98b23"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:21:44.240404Z","signature_b64":"nNxD2ohV8PbDEZkrt8n2ZGhN3rr+A8PD0hcPAeJogWfRDGtbeoGHU+m2iLwHtzFLyZRuxLsG4cSyibvrPlxTDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b12deca2a6188c79a604fa9dd6ccac606bca2d6bbf10dd5ac6b42f2c0d9ffb4c","last_reissued_at":"2026-07-05T11:21:44.239789Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:21:44.239789Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Will Pre-Training Ever End? A First Step Toward Next-Generation Foundation MLLMs via Self-Improving Systematic Cognition","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chengyue Wu, Chi Chen, Da Peng, Helen Meng, Jen-tse Huang, Maosong Sun, Wei Ke, Xiaoying Zhang, Yipeng Zhang, Zonghao Guo","submitted_at":"2025-03-16T00:25:13Z","abstract_excerpt":"Recent progress in (multimodal) large language models ((M)LLMs) has shifted focus from pre-training to inference-time computation and post-training optimization, largely due to concerns over the availability of high-quality human data. However, these strategies alone are insufficient to drive substantial model improvements. We argue that effective model advancement requires strong synergy among pre-training, inference-time computation, and post-training optimization. In this paper, we introduce Self-Improving cognition (SIcog), a self-learning framework for constructing next-generation foundat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.12303","kind":"arxiv","version":6},"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/2503.12303/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2503.12303","created_at":"2026-07-05T11:21:44.239878+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.12303v6","created_at":"2026-07-05T11:21:44.239878+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.12303","created_at":"2026-07-05T11:21:44.239878+00:00"},{"alias_kind":"pith_short_12","alias_value":"WEW6ZIVGDCGH","created_at":"2026-07-05T11:21:44.239878+00:00"},{"alias_kind":"pith_short_16","alias_value":"WEW6ZIVGDCGHTJQE","created_at":"2026-07-05T11:21:44.239878+00:00"},{"alias_kind":"pith_short_8","alias_value":"WEW6ZIVG","created_at":"2026-07-05T11:21:44.239878+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2503.21776","citing_title":"Video-R1: Reinforcing Video Reasoning in MLLMs","ref_index":44,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB","json":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB.json","graph_json":"https://pith.science/api/pith-number/WEW6ZIVGDCGHTJQE7KO5NTFMMB/graph.json","events_json":"https://pith.science/api/pith-number/WEW6ZIVGDCGHTJQE7KO5NTFMMB/events.json","paper":"https://pith.science/paper/WEW6ZIVG"},"agent_actions":{"view_html":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB","download_json":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB.json","view_paper":"https://pith.science/paper/WEW6ZIVG","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.12303&json=true","fetch_graph":"https://pith.science/api/pith-number/WEW6ZIVGDCGHTJQE7KO5NTFMMB/graph.json","fetch_events":"https://pith.science/api/pith-number/WEW6ZIVGDCGHTJQE7KO5NTFMMB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB/action/storage_attestation","attest_author":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB/action/author_attestation","sign_citation":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB/action/citation_signature","submit_replication":"https://pith.science/pith/WEW6ZIVGDCGHTJQE7KO5NTFMMB/action/replication_record"}},"created_at":"2026-07-05T11:21:44.239878+00:00","updated_at":"2026-07-05T11:21:44.239878+00:00"}