{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:H7ESHR5ETQE5DO354J5CHBOYKC","short_pith_number":"pith:H7ESHR5E","schema_version":"1.0","canonical_sha256":"3fc923c7a49c09d1bb7de27a2385d850824f159c82dc61811807091c780221e2","source":{"kind":"arxiv","id":"2412.10028","version":4},"attestation_state":"computed","paper":{"title":"Mr. DETR++: Instructive Multi-Route Training for Detection Transformers with Mixture-of-Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chang-Bin Zhang, Kai Han, Yujie Zhong","submitted_at":"2024-12-13T10:39:27Z","abstract_excerpt":"Existing methods enhance the training of detection transformers by incorporating an auxiliary one-to-many assignment. In this work, we treat the model as a multi-task framework, simultaneously performing one-to-one and one-to-many predictions. We investigate the roles of each component in the transformer decoder across these two training targets, including self-attention, cross-attention, and feed-forward network. Our empirical results demonstrate that any independent component in the decoder can effectively learn both targets simultaneously, even when other components are shared. This finding"},"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":"2412.10028","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2024-12-13T10:39:27Z","cross_cats_sorted":[],"title_canon_sha256":"52f6a3f83baec248b2ce1aeb856731281187a9a4e4bf76504a8dda69297f9d1d","abstract_canon_sha256":"0efff0fc9edbd486618fdf15f569b5752ee6c1e669eae63f6697a84a26716ca0"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:27:22.709400Z","signature_b64":"369Yn+CbcwwRfNkL5h9Vwz8D079mF+eJwl4tIz37/vZaQLhI4wEE67HP/x0zqRdZ8Do8UTDk+aqbPkTj/l6qCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3fc923c7a49c09d1bb7de27a2385d850824f159c82dc61811807091c780221e2","last_reissued_at":"2026-07-05T11:27:22.708857Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:27:22.708857Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Mr. DETR++: Instructive Multi-Route Training for Detection Transformers with Mixture-of-Experts","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chang-Bin Zhang, Kai Han, Yujie Zhong","submitted_at":"2024-12-13T10:39:27Z","abstract_excerpt":"Existing methods enhance the training of detection transformers by incorporating an auxiliary one-to-many assignment. In this work, we treat the model as a multi-task framework, simultaneously performing one-to-one and one-to-many predictions. We investigate the roles of each component in the transformer decoder across these two training targets, including self-attention, cross-attention, and feed-forward network. Our empirical results demonstrate that any independent component in the decoder can effectively learn both targets simultaneously, even when other components are shared. This finding"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.10028","kind":"arxiv","version":4},"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/2412.10028/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":"2412.10028","created_at":"2026-07-05T11:27:22.708925+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.10028v4","created_at":"2026-07-05T11:27:22.708925+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.10028","created_at":"2026-07-05T11:27:22.708925+00:00"},{"alias_kind":"pith_short_12","alias_value":"H7ESHR5ETQE5","created_at":"2026-07-05T11:27:22.708925+00:00"},{"alias_kind":"pith_short_16","alias_value":"H7ESHR5ETQE5DO35","created_at":"2026-07-05T11:27:22.708925+00:00"},{"alias_kind":"pith_short_8","alias_value":"H7ESHR5E","created_at":"2026-07-05T11:27:22.708925+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.04908","citing_title":"HI-MoE: Hierarchical Instance-Conditioned Mixture-of-Experts for Object Detection","ref_index":13,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC","json":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC.json","graph_json":"https://pith.science/api/pith-number/H7ESHR5ETQE5DO354J5CHBOYKC/graph.json","events_json":"https://pith.science/api/pith-number/H7ESHR5ETQE5DO354J5CHBOYKC/events.json","paper":"https://pith.science/paper/H7ESHR5E"},"agent_actions":{"view_html":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC","download_json":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC.json","view_paper":"https://pith.science/paper/H7ESHR5E","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.10028&json=true","fetch_graph":"https://pith.science/api/pith-number/H7ESHR5ETQE5DO354J5CHBOYKC/graph.json","fetch_events":"https://pith.science/api/pith-number/H7ESHR5ETQE5DO354J5CHBOYKC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC/action/storage_attestation","attest_author":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC/action/author_attestation","sign_citation":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC/action/citation_signature","submit_replication":"https://pith.science/pith/H7ESHR5ETQE5DO354J5CHBOYKC/action/replication_record"}},"created_at":"2026-07-05T11:27:22.708925+00:00","updated_at":"2026-07-05T11:27:22.708925+00:00"}