{"id":"https://openalex.org/W7165729524","doi":"https://doi.org/10.48550/arxiv.2606.24447","title":"P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scaling","display_name":"P-MTP: Efficient Document Parsing via Multi-Token Prediction with Progressive Depth Scaling","publication_year":2026,"publication_date":"2026-06-23","ids":{"openalex":"https://openalex.org/W7165729524","doi":"https://doi.org/10.48550/arxiv.2606.24447"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.24447","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.24447","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.24447","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139259845","display_name":"Le Xiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiang, Le","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139302087","display_name":"Chenxi Zhai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhai, Chenxi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139253098","display_name":"Shu Wei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Shu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139270484","display_name":"Jingjing Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Jingjing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100570870","display_name":"Qunyi Xie","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xie, Qunyi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139260383","display_name":"Xiao Tan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tan, Xiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5111571684","display_name":"Kunbin Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Kunbin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139300237","display_name":"Wei He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.4090999960899353,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.4090999960899353,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.17960000038146973,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.10570000112056732,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/parsing","display_name":"Parsing","score":0.6671000123023987},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6340000033378601},{"id":"https://openalex.org/keywords/reliability","display_name":"Reliability (semiconductor)","score":0.5148000121116638},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.45879998803138733},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.43869999051094055},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.4277999997138977},{"id":"https://openalex.org/keywords/noise","display_name":"Noise (video)","score":0.42750000953674316},{"id":"https://openalex.org/keywords/margin","display_name":"Margin (machine learning)","score":0.40070000290870667}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8158000111579895},{"id":"https://openalex.org/C186644900","wikidata":"https://www.wikidata.org/wiki/Q194152","display_name":"Parsing","level":2,"score":0.6671000123023987},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6340000033378601},{"id":"https://openalex.org/C43214815","wikidata":"https://www.wikidata.org/wiki/Q7310987","display_name":"Reliability (semiconductor)","level":3,"score":0.5148000121116638},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.45879998803138733},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4456999897956848},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.43869999051094055},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.4277999997138977},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.42750000953674316},{"id":"https://openalex.org/C774472","wikidata":"https://www.wikidata.org/wiki/Q6760393","display_name":"Margin (machine learning)","level":2,"score":0.40070000290870667},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.388700008392334},{"id":"https://openalex.org/C60690694","wikidata":"https://www.wikidata.org/wiki/Q894902","display_name":"Bottom-up parsing","level":4,"score":0.3741999864578247},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3709999918937683},{"id":"https://openalex.org/C2777735758","wikidata":"https://www.wikidata.org/wiki/Q817765","display_name":"Path (computing)","level":2,"score":0.3637000024318695},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3472000062465668},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3319999873638153},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.3019999861717224},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.29809999465942383},{"id":"https://openalex.org/C2778049539","wikidata":"https://www.wikidata.org/wiki/Q17002908","display_name":"Bayesian optimization","level":2,"score":0.29420000314712524},{"id":"https://openalex.org/C89600930","wikidata":"https://www.wikidata.org/wiki/Q1423946","display_name":"Segmentation","level":2,"score":0.29159998893737793},{"id":"https://openalex.org/C75684735","wikidata":"https://www.wikidata.org/wiki/Q858810","display_name":"Big data","level":2,"score":0.29089999198913574},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.2784000039100647},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.2777000069618225},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.2678000032901764},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.260699987411499}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.24447","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.24447","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.24447","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.24447","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"score":0.45414653420448303,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Vision-Language":[0],"Models":[1],"(VLMs)":[2],"have":[3],"revolutionized":[4],"document":[5,76,188],"parsing":[6,189],"by":[7,42,137],"enabling":[8,113],"end-to-end":[9],"mapping":[10],"from":[11],"images":[12],"to":[13,47,69,116,128,168],"structured":[14],"text,":[15],"imposing":[16],"a":[17,32,57,65,169],"significant":[18],"latency":[19],"bottleneck,":[20],"particularly":[21],"for":[22,35,74],"token-dense":[23],"documents.":[24],"While":[25],"Multi-Token":[26,62],"Prediction":[27],"(MTP)":[28],"has":[29],"emerged":[30],"as":[31],"promising":[33],"approach":[34],"accelerating":[36],"inference,":[37,143],"its":[38],"potential":[39],"is":[40],"constrained":[41],"optimization":[43,111],"instability":[44],"when":[45],"scaling":[46],"deeper":[48],"look-ahead":[49,72,88,120,132,184],"depth.":[50],"In":[51],"this":[52],"paper,":[53],"we":[54,79,123],"propose":[55,124],"\\textbf{P-MTP},":[56],"framework":[58],"that":[59,84,164],"leverages":[60],"\\textbf{Progressive":[61],"Prediction}":[63],"with":[64,172],"lightweight":[66],"MTP":[67,185],"module":[68],"scale":[70],"the":[71,114,130,151,178,187],"depth":[73,133],"high-throughput":[75],"parsing.":[77],"Specifically,":[78],"introduce":[80],"Progressive":[81],"Curriculum":[82],"Loss":[83],"adaptively":[85,138],"re-weights":[86],"different":[87],"depths":[89],"using":[90],"cumulative":[91],"path":[92],"reliability":[93],"and":[94,134,148,161],"retrospective":[95],"target":[96],"consistency.":[97],"By":[98],"effectively":[99],"suppressing":[100],"gradient":[101],"noise":[102],"in":[103,175,186],"long-range":[104],"predictions,":[105],"P-MTP,":[106,165],"facilitates":[107],"an":[108],"automated":[109],"easy-to-hard":[110],"transition,":[112],"model":[115],"master":[117],"increasingly":[118],"distant":[119],"depths.":[121],"Furthermore,":[122],"Confidence-Gated":[125],"Dynamic":[126],"Drafting":[127],"maximize":[129],"effective":[131],"acceptance":[135],"rate":[136],"calibrating":[139],"speculative":[140],"length":[141],"during":[142],"thereby":[144],"minimizing":[145],"computational":[146],"waste":[147],"further":[149],"pushing":[150],"boundaries":[152],"of":[153,182],"inference":[154],"speedup.":[155],"Experimental":[156],"results":[157],"across":[158],"multiple":[159],"benchmarks":[160],"architectures":[162],"demonstrate":[163],"achieves":[166],"up":[167],"$5\\times$":[170],"speedup":[171],"negligible":[173],"loss":[174],"accuracy,":[176],"providing":[177],"first":[179],"successful":[180],"validation":[181],"extensive":[183],"domain.":[190]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-25T00:00:00"}