{"id":"https://openalex.org/W7166833243","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1270","title":"Can LLMs Estimate Student Struggles? Human-AI Difficulty Alignment with Proficiency Simulation for Item Difficulty Prediction","display_name":"Can LLMs Estimate Student Struggles? Human-AI Difficulty Alignment with Proficiency Simulation for Item Difficulty Prediction","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166833243","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1270"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.1270","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1270","pdf_url":"https://aclanthology.org/2026.findings-acl.1270.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.1270.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139731890","display_name":"Ming Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ming Li","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139839178","display_name":"Han Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139819405","display_name":"Yunze Xiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yunze Xiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139762147","display_name":"Jian Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jian Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139837421","display_name":"Hong Jiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hong Jiao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139803879","display_name":"Tianyi Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tianyi Zhou","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.82293329,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"25414","last_page":"25441"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T11902","display_name":"Intelligent Tutoring Systems and Adaptive Learning","score":0.6132000088691711,"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/T11902","display_name":"Intelligent Tutoring Systems and Adaptive Learning","score":0.6132000088691711,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.09560000151395798,"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/T11122","display_name":"Online Learning and Analytics","score":0.06840000301599503,"subfield":{"id":"https://openalex.org/subfields/1706","display_name":"Computer Science Applications"},"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/key","display_name":"Key (lock)","score":0.25949999690055847},{"id":"https://openalex.org/keywords/item-response-theory","display_name":"Item response theory","score":0.2387000024318695}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5435000061988831},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.47049999237060547},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3312999904155731},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.2849999964237213},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.25949999690055847},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.25220000743865967},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.2502000033855438},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.24539999663829803},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.23929999768733978},{"id":"https://openalex.org/C19875794","wikidata":"https://www.wikidata.org/wiki/Q1207340","display_name":"Item response theory","level":3,"score":0.2387000024318695}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.1270","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1270","pdf_url":"https://aclanthology.org/2026.findings-acl.1270.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.1270","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1270","pdf_url":"https://aclanthology.org/2026.findings-acl.1270.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education","score":0.60675448179245}],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166833243.pdf","grobid_xml":"https://content.openalex.org/works/W7166833243.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Accurate":[0],"estimation":[1],"of":[2,39,50,83,111,128,149,156],"item":[3],"(question":[4],"or":[5],"task)":[6],"difficulty":[7,101,162],"is":[8,78],"critical":[9,126],"for":[10,54,160],"educational":[11],"assessment":[12],"but":[13],"suffers":[14],"from":[15],"the":[16,36,108,154],"cold":[17],"start":[18],"problem.While":[19],"Large":[20],"Language":[21],"Models":[22],"demonstrate":[23],"superhuman":[24],"problem-solving":[25,142],"capabilities,":[26],"it":[27],"remains":[28],"an":[29,147],"open":[30],"question":[31],"whether":[32],"they":[33],"can":[34],"perceive":[35],"cognitive":[37,151],"struggles":[38],"human":[40,150],"learners.In":[41],"this":[42],"work,":[43],"we":[44,123],"present":[45],"a":[46,70,90,125],"large-scale":[47],"empirical":[48],"analysis":[49],"Human-AI":[51],"Difficulty":[52],"Alignment":[53],"over":[55],"20":[56],"models":[57,87,104,131,159],"across":[58],"diverse":[59],"domains":[60],"such":[61],"as":[62,103,130],"medical":[63],"knowledge":[64],"and":[65],"mathematical":[66],"reasoning.Our":[67],"findings":[68],"reveal":[69],"systematic":[71],"misalignment":[72],"where":[73],"scaling":[74],"up":[75],"model":[76],"size":[77],"not":[79,145],"reliably":[80],"helpful;":[81],"instead":[82],"aligning":[84],"with":[85],"humans,":[86],"converge":[88],"toward":[89],"shared":[91],"machine":[92],"consensus.We":[93],"observe":[94],"that":[95,140],"high":[96],"performance":[97],"often":[98],"impedes":[99],"accurate":[100],"estimation,":[102],"struggle":[105],"to":[106,118,133],"simulate":[107],"capability":[109,143],"limitations":[110],"students":[112],"even":[113],"when":[114],"being":[115],"explicitly":[116],"prompted":[117],"adopt":[119],"specific":[120],"proficiency":[121],"levels.Furthermore,":[122],"identify":[124],"lack":[127],"introspection,":[129],"fail":[132],"predict":[134],"their":[135],"own":[136],"limitations.These":[137],"results":[138],"suggest":[139],"general":[141],"does":[144],"imply":[146],"understanding":[148],"struggles,":[152],"highlighting":[153],"challenge":[155],"using":[157],"current":[158],"automated":[161],"prediction.":[163]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}