{"id":"https://openalex.org/W4400266577","doi":"https://doi.org/10.1093/jamia/ocae090","title":"<i>The first step is the hardest</i>: pitfalls of representing and tokenizing temporal data for large language models","display_name":"<i>The first step is the hardest</i>: pitfalls of representing and tokenizing temporal data for large language models","publication_year":2024,"publication_date":"2024-07-01","ids":{"openalex":"https://openalex.org/W4400266577","doi":"https://doi.org/10.1093/jamia/ocae090","pmid":"https://pubmed.ncbi.nlm.nih.gov/38950417"},"language":"en","primary_location":{"id":"doi:10.1093/jamia/ocae090","is_oa":false,"landing_page_url":"https://doi.org/10.1093/jamia/ocae090","pdf_url":null,"source":{"id":"https://openalex.org/S129839026","display_name":"Journal of the American Medical Informatics Association","issn_l":"1067-5027","issn":["1067-5027","1527-974X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310311648","host_organization_name":"Oxford University Press","host_organization_lineage":["https://openalex.org/P4310311648","https://openalex.org/P4310311647"],"host_organization_lineage_names":["Oxford University Press","University of Oxford"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Journal of the American Medical Informatics Association","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","pubmed"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"http://eprints.gla.ac.uk/view/author/64762.html>","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5030362528","display_name":"Dimitris Spathis","orcid":"https://orcid.org/0000-0001-9761-951X"},"institutions":[{"id":"https://openalex.org/I241749","display_name":"University of Cambridge","ror":"https://ror.org/013meh722","country_code":"GB","type":"education","lineage":["https://openalex.org/I241749"]},{"id":"https://openalex.org/I4210098141","display_name":"Nokia (United Kingdom)","ror":"https://ror.org/00zpf0626","country_code":"GB","type":"company","lineage":["https://openalex.org/I2738502077","https://openalex.org/I4210098141"]}],"countries":["GB"],"is_corresponding":true,"raw_author_name":"Dimitris Spathis","raw_affiliation_strings":["Department of Computer Science and Technology, University of Cambridge , Cambridge, CB3 0FD, United Kingdom","Nokia Bell Labs , Cambridge, CB3 0FA, United Kingdom"],"raw_orcid":"https://orcid.org/0000-0001-9761-951X","affiliations":[{"raw_affiliation_string":"Department of Computer Science and Technology, University of Cambridge , Cambridge, CB3 0FD, United Kingdom","institution_ids":["https://openalex.org/I241749"]},{"raw_affiliation_string":"Nokia Bell Labs , Cambridge, CB3 0FA, United Kingdom","institution_ids":["https://openalex.org/I4210098141"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5053438231","display_name":"Fahim Kawsar","orcid":"https://orcid.org/0000-0001-5057-9557"},"institutions":[{"id":"https://openalex.org/I4210098141","display_name":"Nokia (United Kingdom)","ror":"https://ror.org/00zpf0626","country_code":"GB","type":"company","lineage":["https://openalex.org/I2738502077","https://openalex.org/I4210098141"]},{"id":"https://openalex.org/I7882870","display_name":"University of Glasgow","ror":"https://ror.org/00vtgdb53","country_code":"GB","type":"education","lineage":["https://openalex.org/I7882870"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Fahim Kawsar","raw_affiliation_strings":["Nokia Bell Labs , Cambridge, CB3 0FA, United Kingdom","School of Computing Science, University of Glasgow , Glasgow, G12 8RZ, United Kingdom"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nokia Bell Labs , Cambridge, CB3 0FA, United Kingdom","institution_ids":["https://openalex.org/I4210098141"]},{"raw_affiliation_string":"School of Computing Science, University of Glasgow , Glasgow, G12 8RZ, United Kingdom","institution_ids":["https://openalex.org/I7882870"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":["https://openalex.org/A5030362528"],"corresponding_institution_ids":["https://openalex.org/I241749","https://openalex.org/I4210098141"],"apc_list":{"value":3967,"currency":"USD","value_usd":3967},"apc_paid":null,"fwci":5.9854,"has_fulltext":true,"cited_by_count":25,"citation_normalized_percentile":{"value":0.96703156,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":96,"max":100},"biblio":{"volume":"31","issue":"9","first_page":"2151","last_page":"2158"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.27649998664855957,"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/T10028","display_name":"Topic Modeling","score":0.27649998664855957,"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/T11636","display_name":"Artificial Intelligence in Healthcare and Education","score":0.17090000212192535,"subfield":{"id":"https://openalex.org/subfields/2718","display_name":"Health Informatics"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}},{"id":"https://openalex.org/T11714","display_name":"Multimodal Machine Learning Applications","score":0.133200004696846,"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/computer-science","display_name":"Computer science","score":0.6034125089645386},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.3753175735473633},{"id":"https://openalex.org/keywords/data-science","display_name":"Data science","score":0.3368736505508423},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.32032984495162964}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6034125089645386},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3753175735473633},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.3368736505508423},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.32032984495162964}],"mesh":[{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D006801","descriptor_name":"Humans","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D009323","descriptor_name":"Natural Language Processing","qualifier_ui":null,"qualifier_name":null,"is_major_topic":true},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false},{"descriptor_ui":"D057286","descriptor_name":"Electronic Health Records","qualifier_ui":null,"qualifier_name":null,"is_major_topic":false}],"locations_count":4,"locations":[{"id":"doi:10.1093/jamia/ocae090","is_oa":false,"landing_page_url":"https://doi.org/10.1093/jamia/ocae090","pdf_url":null,"source":{"id":"https://openalex.org/S129839026","display_name":"Journal of the American Medical Informatics Association","issn_l":"1067-5027","issn":["1067-5027","1527-974X"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310311648","host_organization_name":"Oxford University Press","host_organization_lineage":["https://openalex.org/P4310311648","https://openalex.org/P4310311647"],"host_organization_lineage_names":["Oxford University Press","University of Oxford"],"type":"journal"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Journal of the American Medical Informatics Association","raw_type":"journal-article"},{"id":"pmid:38950417","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/38950417","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Journal of the American Medical Informatics Association : JAMIA","raw_type":null},{"id":"pmh:oai:eprints.gla.ac.uk:329873","is_oa":true,"landing_page_url":"http://eprints.gla.ac.uk/view/author/64762.html>","pdf_url":null,"source":{"id":"https://openalex.org/S4210235606","display_name":"ENLIGHTEN (Jurnal Bimbingan dan Konseling Islam)","issn_l":"2622-8912","issn":["2622-8912","2622-8920"],"is_oa":true,"is_in_doaj":true,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"acceptedVersion","is_accepted":true,"is_published":false,"raw_source_name":null,"raw_type":"Articles"},{"id":"pmh:oai:pubmedcentral.nih.gov:11339515","is_oa":true,"landing_page_url":"https://www.ncbi.nlm.nih.gov/pmc/articles/11339515","pdf_url":"https://pmc.ncbi.nlm.nih.gov/articles/PMC11339515/pdf/ocae090.pdf","source":{"id":"https://openalex.org/S2764455111","display_name":"PubMed Central","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"J Am Med Inform Assoc","raw_type":"Text"}],"best_oa_location":{"id":"pmh:oai:eprints.gla.ac.uk:329873","is_oa":true,"landing_page_url":"http://eprints.gla.ac.uk/view/author/64762.html>","pdf_url":null,"source":{"id":"https://openalex.org/S4210235606","display_name":"ENLIGHTEN (Jurnal Bimbingan dan Konseling Islam)","issn_l":"2622-8912","issn":["2622-8912","2622-8920"],"is_oa":true,"is_in_doaj":true,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"journal"},"license":"cc-by-sa","license_id":"https://openalex.org/licenses/cc-by-sa","version":"acceptedVersion","is_accepted":true,"is_published":false,"raw_source_name":null,"raw_type":"Articles"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","score":0.5699999928474426,"display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":15,"referenced_works":["https://openalex.org/W2972660824","https://openalex.org/W3099025572","https://openalex.org/W3128981305","https://openalex.org/W3185341429","https://openalex.org/W4211062518","https://openalex.org/W4225432580","https://openalex.org/W4283031227","https://openalex.org/W4292779060","https://openalex.org/W4294891660","https://openalex.org/W4322766882","https://openalex.org/W4379537722","https://openalex.org/W4394757986","https://openalex.org/W6778883912","https://openalex.org/W6810452849","https://openalex.org/W6838540236"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052","https://openalex.org/W2382290278","https://openalex.org/W3204019825"],"abstract_inverted_index":{"OBJECTIVES:":[0],"Large":[1],"language":[2],"models":[3,221,237,336],"(LLMs)":[4],"have":[5],"demonstrated":[6],"remarkable":[7],"generalization":[8],"and":[9,72,79,87,99,121,163,176,199,287,298],"across":[10],"diverse":[11],"tasks,":[12,155],"leading":[13],"individuals":[14],"to":[15,23,68,75,109,115,202],"increasingly":[16],"use":[17],"them":[18],"as":[19,41,84,303],"personal":[20],"assistants":[21],"due":[22,114],"their":[24,54,89],"emerging":[25],"reasoning":[26],"capabilities.":[27],"Nevertheless,":[28],"a":[29,128,133],"notable":[30],"obstacle":[31],"emerges":[32],"when":[33,290],"including":[34],"numerical/temporal":[35],"data":[36,42,150,318],"into":[37,60,151,179],"these":[38],"prompts,":[39],"such":[40],"sourced":[43],"from":[44,238],"wearables":[45],"or":[46],"electronic":[47],"health":[48],"records.":[49],"LLMs":[50,110,152,173,252,313],"employ":[51],"tokenizers":[52,64,264,300],"in":[53,277,331],"input":[55,321],"that":[56,105,147,171,191,206,309,327],"break":[57],"down":[58],"text":[59],"smaller":[61],"units.":[62],"However,":[63],"are":[65,265,301],"not":[66,243,266],"designed":[67],"represent":[69],"numerical":[70,269],"values":[71,83,178],"might":[73],"struggle":[74,185],"understand":[76],"repetitive":[77],"patterns":[78],"context,":[80],"treating":[81],"consecutive":[82],"separate":[85],"tokens":[86],"disregarding":[88],"temporal":[90,101,187,259,292,317],"relationships.":[91],"This":[92],"article":[93],"discusses":[94],"the":[95,116,139,207,218,232,246,272,320,340,346],"challenges":[96],"of":[97,220,231,251,274,342,348],"representing":[98],"tokenizing":[100,132],"data.":[102,188,260,293],"It":[103],"argues":[104],"naively":[106],"passing":[107],"timeseries":[108,149,200,275],"can":[111],"be":[112],"ineffective":[113],"modality":[117],"gap":[118],"between":[119],"numbers":[120],"text.":[122],"MATERIALS":[123],"AND":[124],"METHODS:":[125],"We":[126,142,189,262,282,307,325],"conduct":[127],"case":[129,168],"study":[130,169],"by":[131],"sample":[134],"mobile":[135],"sensing":[136],"dataset":[137],"using":[138],"OpenAI":[140],"tokenizer.":[141],"also":[143],"review":[144],"recent":[145],"works":[146,193],"feed":[148],"for":[153,223,258,268],"human-centric":[154],"outlining":[156],"common":[157],"experimental":[158],"setups":[159],"like":[160],"zero-shot":[161,249],"prompting":[162,250],"few-shot":[164],"learning.":[165],"RESULTS:":[166],"The":[167,212],"shows":[170],"popular":[172],"split":[174],"timestamps":[175],"sensor":[177],"multiple":[180],"nonmeaningful":[181],"tokens,":[182],"indicating":[183],"they":[184],"with":[186],"find":[190],"preliminary":[192],"rely":[194],"heavily":[195],"on":[196],"prompt":[197],"engineering":[198],"aggregation":[201],"\"ground\"":[203],"LLMs,":[204],"hinting":[205],"\"modality":[208],"gap\"":[209],"hampers":[210],"progress.":[211],"literature":[213],"was":[214],"critically":[215],"analyzed":[216],"through":[217],"lens":[219],"optimizing":[222],"expressiveness":[224,257,286],"versus":[225],"parameter":[226],"efficiency.":[227],"On":[228,245],"one":[229],"end":[230],"spectrum,":[233],"training":[234,278],"large":[235],"domain-specific":[236],"scratch":[239],"is":[240,253,323],"expressive":[241],"but":[242,255],"parameter-efficient.":[244],"other":[247],"end,":[248],"parameter-efficient":[254],"lacks":[256],"DISCUSSION:":[261],"argue":[263,326],"optimized":[267],"data,":[270],"while":[271],"scarcity":[273],"examples":[276],"corpora":[279],"exacerbates":[280],"difficulties.":[281],"advocate":[283],"balancing":[284],"model":[285,296],"computational":[288],"efficiency":[289],"integrating":[291],"Prompt":[294],"tuning,":[295],"grafting,":[297],"improved":[299],"highlighted":[302],"promising":[304,311],"directions.":[305],"CONCLUSION:":[306],"underscore":[308],"despite":[310],"capabilities,":[312],"cannot":[314],"meaningfully":[315],"process":[316],"unless":[319],"representation":[322],"addressed.":[324],"this":[328],"paradigm":[329],"shift":[330],"how":[332],"we":[333],"leverage":[334],"pretrained":[335],"will":[337],"particularly":[338],"affect":[339],"area":[341],"biomedical":[343],"signals,":[344],"given":[345],"lack":[347],"modality-specific":[349],"foundation":[350],"models.":[351]},"counts_by_year":[{"year":2026,"cited_by_count":7},{"year":2025,"cited_by_count":15},{"year":2024,"cited_by_count":3}],"updated_date":"2026-08-01T09:00:35.917206","created_date":"2024-07-03T00:00:00"}