{"id":"https://openalex.org/W7164054534","doi":"https://doi.org/10.48550/arxiv.2606.08578","title":"Lost in the Non-convex Loss Landscape: How to Fine-tune the Large Time Series Model?","display_name":"Lost in the Non-convex Loss Landscape: How to Fine-tune the Large Time Series Model?","publication_year":2026,"publication_date":"2026-06-07","ids":{"openalex":"https://openalex.org/W7164054534","doi":"https://doi.org/10.48550/arxiv.2606.08578"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.08578","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08578","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.2606.08578","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138267204","display_name":"Xu Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Xu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138269664","display_name":"Peang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Peang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138249132","display_name":"Wei Wang (17594)","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, 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/T13702","display_name":"Machine Learning in Healthcare","score":0.47760000824928284,"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"}},"topics":[{"id":"https://openalex.org/T13702","display_name":"Machine Learning in Healthcare","score":0.47760000824928284,"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/T12205","display_name":"Time Series Analysis and Forecasting","score":0.2362000048160553,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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":0.04820000007748604,"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/overfitting","display_name":"Overfitting","score":0.8295999765396118},{"id":"https://openalex.org/keywords/initialization","display_name":"Initialization","score":0.7592999935150146},{"id":"https://openalex.org/keywords/series","display_name":"Series (stratigraphy)","score":0.6581000089645386},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.6363999843597412},{"id":"https://openalex.org/keywords/maxima-and-minima","display_name":"Maxima and minima","score":0.6238999962806702},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5604000091552734},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.5557000041007996},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.4390000104904175}],"concepts":[{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.8295999765396118},{"id":"https://openalex.org/C114466953","wikidata":"https://www.wikidata.org/wiki/Q6034165","display_name":"Initialization","level":2,"score":0.7592999935150146},{"id":"https://openalex.org/C143724316","wikidata":"https://www.wikidata.org/wiki/Q312468","display_name":"Series (stratigraphy)","level":2,"score":0.6581000089645386},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.6363999843597412},{"id":"https://openalex.org/C186633575","wikidata":"https://www.wikidata.org/wiki/Q845060","display_name":"Maxima and minima","level":2,"score":0.6238999962806702},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5881999731063843},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5604000091552734},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5557000041007996},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.4390000104904175},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.4296000003814697},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.3788999915122986},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.37790000438690186},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.376800000667572},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.33730000257492065},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.328900009393692},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.32330000400543213},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3188999891281128},{"id":"https://openalex.org/C151406439","wikidata":"https://www.wikidata.org/wiki/Q186588","display_name":"Time series","level":2,"score":0.28360000252723694},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2818000018596649},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.2777000069618225},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.25619998574256897},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.2526000142097473},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.25099998712539673}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.08578","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08578","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.2606.08578","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08578","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":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recently,":[0],"large":[1,15],"time":[2],"series":[3],"models":[4],"(LTSMs)":[5],"have":[6],"gained":[7],"increasing":[8],"attention":[9],"due":[10],"to":[11,14,46,55,97,115],"their":[12],"similarities":[13],"language":[16],"models,":[17],"including":[18,171],"flexible":[19],"context":[20],"length,":[21],"scalability,":[22],"and":[23,58,103,155,179],"task":[24],"generality,":[25],"outperforming":[26],"advanced":[27],"task-specific":[28],"models.":[29],"However,":[30],"prior":[31],"studies":[32],"indicate":[33],"that":[34],"pre-trained":[35,113,126],"LTSMs":[36],"may":[37],"exhibit":[38],"a":[39,50,84,99],"poorly":[40],"conditioned":[41],"non-convex":[42],"loss":[43,101],"landscape,":[44,102],"leading":[45],"limited":[47],"trainability.":[48],"As":[49],"result,":[51],"direct":[52],"fine-tuning":[53,86],"tends":[54],"cause":[56],"overfitting":[57],"suboptimal":[59],"performance,":[60],"sometimes":[61],"even":[62],"worse":[63],"than":[64],"training":[65],"from":[66,149],"scratch,":[67],"substantially":[68],"diminishing":[69],"the":[70,112,117,190],"benefits":[71],"of":[72,111],"pre-training.":[73],"To":[74],"overcome":[75],"this":[76],"limitation,":[77],"we":[78,89],"propose":[79],"Smoothed":[80],"Full":[81],"Fine-tuning":[82],"(SFF),":[83],"novel":[85],"technology.":[87],"Specifically,":[88],"construct":[90],"an":[91,135],"auxiliary":[92],"LTSM":[93],"via":[94],"random":[95],"initialization":[96],"obtain":[98],"smoother":[100,154],"then":[104],"linearly":[105],"interpolate":[106],"its":[107],"weights":[108],"with":[109],"those":[110],"model":[114],"smooth":[116],"original":[118],"landscape.":[119],"This":[120],"process":[121],"improves":[122],"trainability":[123],"while":[124],"preserving":[125],"knowledge,":[127],"thereby":[128],"enabling":[129],"more":[130,156],"effective":[131],"downstream":[132,183],"fine-tuning.":[133],"From":[134],"optimization":[136],"perspective,":[137],"SFF":[138],"perturbs":[139],"sharp":[140],"minima":[141],"without":[142],"significantly":[143],"harming":[144],"flat":[145],"regions,":[146],"facilitating":[147],"escape":[148],"poor":[150],"local":[151],"basins":[152],"toward":[153],"generalizable":[157],"solutions.":[158],"Extensive":[159],"experiments":[160],"on":[161,181],"benchmark":[162],"datasets":[163],"demonstrate":[164],"consistent":[165],"improvements":[166],"across":[167],"eight":[168],"representative":[169],"LTSMs,":[170],"Timer,":[172],"TimesFM,":[173],"MOMENT,":[174],"UniTS,":[175],"MOIRAI,":[176],"Chronos,":[177],"TTMs,":[178],"Sundial,":[180],"diverse":[182],"tasks.":[184],"The":[185],"code":[186],"is":[187],"available":[188],"at":[189],"link:":[191],"https://github.com/Meteor-Stars/SFF.":[192]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-10T00:00:00"}