{"id":"https://openalex.org/W4416377937","doi":"https://doi.org/10.48550/arxiv.2503.16929","title":"TEMPLE: Incentivizing Temporal Understanding of Video Large Language Models via Progressive Pre-SFT Alignment","display_name":"TEMPLE: Incentivizing Temporal Understanding of Video Large Language Models via Progressive Pre-SFT Alignment","publication_year":2025,"publication_date":"2025-03-21","ids":{"openalex":"https://openalex.org/W4416377937","doi":"https://doi.org/10.48550/arxiv.2503.16929"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2503.16929","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.16929","pdf_url":"https://arxiv.org/pdf/2503.16929","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2503.16929","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100723941","display_name":"Shicheng Li","orcid":"https://orcid.org/0000-0002-5724-0641"},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Li, Shicheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100440326","display_name":"Lei Li","orcid":"https://orcid.org/0000-0002-5374-7293"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Lei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5109515733","display_name":"Kun Ouyang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ouyang, Kun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101030723","display_name":"Shuhuai Ren","orcid":"https://orcid.org/0009-0001-9998-864X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ren, Shuhuai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5044541656","display_name":"Yuanxin Liu","orcid":"https://orcid.org/0000-0002-0673-4887"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yuanxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020022791","display_name":"Yuanxing Zhang","orcid":"https://orcid.org/0000-0003-1460-8124"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Yuanxing","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101586838","display_name":"Fuzheng Zhang","orcid":"https://orcid.org/0000-0001-6232-8464"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Fuzheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101497114","display_name":"Lingpeng Kong","orcid":"https://orcid.org/0000-0002-9407-7193"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kong, Lingpeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100453144","display_name":"Qi Liu","orcid":"https://orcid.org/0000-0001-5378-6404"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Qi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5006497264","display_name":"Xu Sun","orcid":"https://orcid.org/0000-0001-9155-0427"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Xu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":10,"corresponding_author_ids":["https://openalex.org/A5100723941"],"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":0.7153000235557556,"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":0.7153000235557556,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.06589999794960022,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.055799998342990875,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.5508000254631042},{"id":"https://openalex.org/keywords/complement","display_name":"Complement (music)","score":0.5001999735832214},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4708999991416931},{"id":"https://openalex.org/keywords/preference-learning","display_name":"Preference learning","score":0.46720001101493835},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.4575999975204468},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.45489999651908875},{"id":"https://openalex.org/keywords/preference","display_name":"Preference","score":0.4318000078201294}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7645000219345093},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5874999761581421},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.5508000254631042},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5116999745368958},{"id":"https://openalex.org/C112313634","wikidata":"https://www.wikidata.org/wiki/Q7886648","display_name":"Complement (music)","level":5,"score":0.5001999735832214},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4708999991416931},{"id":"https://openalex.org/C181204326","wikidata":"https://www.wikidata.org/wiki/Q7239820","display_name":"Preference learning","level":3,"score":0.46720001101493835},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.4575999975204468},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.45489999651908875},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.4318000078201294},{"id":"https://openalex.org/C174348530","wikidata":"https://www.wikidata.org/wiki/Q188635","display_name":"Bridging (networking)","level":2,"score":0.3490999937057495},{"id":"https://openalex.org/C77277458","wikidata":"https://www.wikidata.org/wiki/Q1969246","display_name":"Temporal database","level":2,"score":0.34360000491142273},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.3206999897956848},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.2919999957084656},{"id":"https://openalex.org/C58489278","wikidata":"https://www.wikidata.org/wiki/Q1172284","display_name":"Data set","level":2,"score":0.29089999198913574},{"id":"https://openalex.org/C132010649","wikidata":"https://www.wikidata.org/wiki/Q189222","display_name":"Intuition","level":2,"score":0.259799987077713},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2587999999523163},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.25130000710487366}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2503.16929","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.16929","pdf_url":"https://arxiv.org/pdf/2503.16929","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2503.16929","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2503.16929","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2503.16929","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.16929","pdf_url":"https://arxiv.org/pdf/2503.16929","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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"awards":[{"id":"https://openalex.org/G4370721888","display_name":null,"funder_award_id":"62176002","funder_id":"https://openalex.org/F4320321001","funder_display_name":"National Natural Science Foundation of China"}],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Video":[0,170,204],"Large":[1],"Language":[2],"Models":[3],"(Video":[4],"LLMs)":[5],"have":[6],"achieved":[7],"significant":[8],"success":[9],"by":[10,18],"adopting":[11],"the":[12,35,40,48,199],"paradigm":[13],"of":[14,181],"large-scale":[15],"pre-training":[16],"followed":[17],"supervised":[19],"fine-tuning":[20],"(SFT).":[21],"However,":[22],"existing":[23],"approaches":[24],"struggle":[25],"with":[26,176],"temporal":[27,32,50,67,77,160],"reasoning":[28,68],"due":[29],"to":[30,146,157,195],"weak":[31],"correspondence":[33],"in":[34,47,80],"data":[36,115,148],"and":[37,104,110,125,150,192],"over-reliance":[38],"on":[39,108],"next-token":[41],"prediction":[42],"paradigm},":[43],"which":[44,141],"collectively":[45],"result":[46],"absence":[49],"supervision.":[51],"To":[52,75],"address":[53,76],"these":[54],"limitations,":[55],"we":[56,82,117],"propose":[57,126],"TEMPLE":[58,188],"(TEMporal":[59],"Preference":[60,72],"LEarning),":[61],"a":[62,127,137,177,190],"systematic":[63],"framework":[64],"that":[65,165],"enhances":[66],"capabilities":[69],"through":[70],"Direct":[71],"Optimization":[73],"(DPO).":[74],"information":[78],"scarcity":[79],"data,":[81],"introduce":[83],"an":[84],"automated":[85],"pipeline":[86],"for":[87,201],"systematically":[88],"constructing":[89],"temporality-intensive":[90],"preference":[91,123,152],"pairs":[92],"comprising":[93],"three":[94],"steps:":[95],"selecting":[96],"temporally":[97],"rich":[98],"videos,":[99],"designing":[100],"video-specific":[101],"perturbation":[102,144],"strategies,":[103],"evaluating":[105],"model":[106],"responses":[107],"clean":[109],"perturbed":[111],"inputs.":[112],"Complementing":[113],"this":[114],"pipeline,":[116],"provide":[118],"additional":[119],"supervision":[120],"signals":[121],"via":[122],"learning":[124,139],"novel":[128],"Progressive":[129],"Pre-SFT":[130],"Alignment":[131],"strategy":[132,140],"featuring":[133],"two":[134],"key":[135],"innovations:":[136],"curriculum":[138],"progressively":[142],"increases":[143],"difficulty":[145],"maximize":[147],"efficiency;":[149],"applying":[151],"optimization":[153],"before":[154],"instruction":[155],"tuning":[156],"incentivize":[158],"fundamental":[159],"alignment.":[161],"Extensive":[162],"experiments":[163],"demonstrate":[164],"our":[166],"approach":[167],"consistently":[168],"improves":[169],"LLM":[171],"performance":[172],"across":[173],"multiple":[174],"benchmarks":[175],"relatively":[178],"small":[179],"set":[180],"self-generated":[182],"DPO":[183],"data.":[184],"Our":[185],"findings":[186],"highlight":[187],"as":[189],"scalable":[191],"efficient":[193],"complement":[194],"SFT-based":[196],"methods,":[197],"paving":[198],"way":[200],"developing":[202],"reliable":[203],"LLMs.":[205]},"counts_by_year":[],"updated_date":"2026-05-08T15:41:06.802602","created_date":"2025-10-10T00:00:00"}