{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:XQQCCYU7VGKYKGKJXCJ64X7LJ7","short_pith_number":"pith:XQQCCYU7","schema_version":"1.0","canonical_sha256":"bc2021629fa995851949b893ee5feb4fd2f2ac1009cabf8418c4e6636acf73b8","source":{"kind":"arxiv","id":"2606.24447","version":1},"attestation_state":"computed","paper":{"title":"P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenxi Zhai, Jingjing Wu, Kunbin Chen, Le Xiang, Qunyi Xie, Shu Wei, Wei He, Xiao Tan","submitted_at":"2026-06-23T11:34:28Z","abstract_excerpt":"Vision-Language Models (VLMs) have revolutionized document parsing by enabling end-to-end mapping from images to structured text, imposing a significant latency bottleneck, particularly for token-dense documents. While Multi-Token Prediction (MTP) has emerged as a promising approach for accelerating inference, its potential is constrained by optimization instability when scaling to deeper look-ahead depth. In this paper, we propose \\textbf{P-MTP}, a framework that leverages \\textbf{Progressive Multi-Token Prediction} with a lightweight MTP module to scale the look-ahead depth for high-throughp"},"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":"2606.24447","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-06-23T11:34:28Z","cross_cats_sorted":[],"title_canon_sha256":"eaf5647196fa2c5d178ecf79dab5b1ce15ec4db95abf3780fa2759b7200bf7b6","abstract_canon_sha256":"15f550f5128a41995973657342132eef40954c7633906189a1f305994ea1584b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-24T01:15:30.657577Z","signature_b64":"CP6nX9aTZ/vX5sh3XQ1fzSbfoZiwSCg7cggxIzvGAdhTpHvHiM6gI7D+re45XGxI8gfqvjD2vUt9PbuxbHoOCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bc2021629fa995851949b893ee5feb4fd2f2ac1009cabf8418c4e6636acf73b8","last_reissued_at":"2026-06-24T01:15:30.657217Z","signature_status":"signed_v1","first_computed_at":"2026-06-24T01:15:30.657217Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scaling","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Chenxi Zhai, Jingjing Wu, Kunbin Chen, Le Xiang, Qunyi Xie, Shu Wei, Wei He, Xiao Tan","submitted_at":"2026-06-23T11:34:28Z","abstract_excerpt":"Vision-Language Models (VLMs) have revolutionized document parsing by enabling end-to-end mapping from images to structured text, imposing a significant latency bottleneck, particularly for token-dense documents. While Multi-Token Prediction (MTP) has emerged as a promising approach for accelerating inference, its potential is constrained by optimization instability when scaling to deeper look-ahead depth. In this paper, we propose \\textbf{P-MTP}, a framework that leverages \\textbf{Progressive Multi-Token Prediction} with a lightweight MTP module to scale the look-ahead depth for high-throughp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.24447","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/2606.24447/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":"2606.24447","created_at":"2026-06-24T01:15:30.657273+00:00"},{"alias_kind":"arxiv_version","alias_value":"2606.24447v1","created_at":"2026-06-24T01:15:30.657273+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.24447","created_at":"2026-06-24T01:15:30.657273+00:00"},{"alias_kind":"pith_short_12","alias_value":"XQQCCYU7VGKY","created_at":"2026-06-24T01:15:30.657273+00:00"},{"alias_kind":"pith_short_16","alias_value":"XQQCCYU7VGKYKGKJ","created_at":"2026-06-24T01:15:30.657273+00:00"},{"alias_kind":"pith_short_8","alias_value":"XQQCCYU7","created_at":"2026-06-24T01:15:30.657273+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7","json":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7.json","graph_json":"https://pith.science/api/pith-number/XQQCCYU7VGKYKGKJXCJ64X7LJ7/graph.json","events_json":"https://pith.science/api/pith-number/XQQCCYU7VGKYKGKJXCJ64X7LJ7/events.json","paper":"https://pith.science/paper/XQQCCYU7"},"agent_actions":{"view_html":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7","download_json":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7.json","view_paper":"https://pith.science/paper/XQQCCYU7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2606.24447&json=true","fetch_graph":"https://pith.science/api/pith-number/XQQCCYU7VGKYKGKJXCJ64X7LJ7/graph.json","fetch_events":"https://pith.science/api/pith-number/XQQCCYU7VGKYKGKJXCJ64X7LJ7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7/action/storage_attestation","attest_author":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7/action/author_attestation","sign_citation":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7/action/citation_signature","submit_replication":"https://pith.science/pith/XQQCCYU7VGKYKGKJXCJ64X7LJ7/action/replication_record"}},"created_at":"2026-06-24T01:15:30.657273+00:00","updated_at":"2026-06-24T01:15:30.657273+00:00"}