{"id":"https://openalex.org/W7165803349","doi":"https://doi.org/10.48550/arxiv.2606.24839","title":"Grading the Grader: Lessons from Evaluating an Agentic Data Analysis System","display_name":"Grading the Grader: Lessons from Evaluating an Agentic Data Analysis System","publication_year":2026,"publication_date":"2026-06-23","ids":{"openalex":"https://openalex.org/W7165803349","doi":"https://doi.org/10.48550/arxiv.2606.24839"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.24839","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.24839","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":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.2606.24839","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139300189","display_name":"Tian Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Tian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139240301","display_name":"Kai-Tai Hsu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hsu, Kai-Tai","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/T10028","display_name":"Topic Modeling","score":0.3310000002384186,"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":0.3310000002384186,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.1932000070810318,"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/T13629","display_name":"Text Readability and Simplification","score":0.05979999899864197,"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/grading","display_name":"Grading (engineering)","score":0.8183000087738037},{"id":"https://openalex.org/keywords/recall","display_name":"Recall","score":0.5580000281333923},{"id":"https://openalex.org/keywords/precision-and-recall","display_name":"Precision and recall","score":0.49540001153945923},{"id":"https://openalex.org/keywords/metadata","display_name":"Metadata","score":0.436599999666214},{"id":"https://openalex.org/keywords/data-quality","display_name":"Data quality","score":0.3977999985218048},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.32519999146461487},{"id":"https://openalex.org/keywords/task-analysis","display_name":"Task analysis","score":0.29660001397132874},{"id":"https://openalex.org/keywords/data-extraction","display_name":"Data extraction","score":0.29100000858306885}],"concepts":[{"id":"https://openalex.org/C2777286243","wikidata":"https://www.wikidata.org/wiki/Q5591926","display_name":"Grading (engineering)","level":2,"score":0.8183000087738037},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6650000214576721},{"id":"https://openalex.org/C100660578","wikidata":"https://www.wikidata.org/wiki/Q18733","display_name":"Recall","level":2,"score":0.5580000281333923},{"id":"https://openalex.org/C81669768","wikidata":"https://www.wikidata.org/wiki/Q2359161","display_name":"Precision and recall","level":2,"score":0.49540001153945923},{"id":"https://openalex.org/C93518851","wikidata":"https://www.wikidata.org/wiki/Q180160","display_name":"Metadata","level":2,"score":0.436599999666214},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.41589999198913574},{"id":"https://openalex.org/C24756922","wikidata":"https://www.wikidata.org/wiki/Q1757694","display_name":"Data quality","level":3,"score":0.3977999985218048},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.38839998841285706},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3822999894618988},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.3564999997615814},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.32519999146461487},{"id":"https://openalex.org/C2522767166","wikidata":"https://www.wikidata.org/wiki/Q2374463","display_name":"Data science","level":1,"score":0.3066999912261963},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.30000001192092896},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.29660001397132874},{"id":"https://openalex.org/C2777466982","wikidata":"https://www.wikidata.org/wiki/Q5227287","display_name":"Data extraction","level":3,"score":0.29100000858306885},{"id":"https://openalex.org/C154690210","wikidata":"https://www.wikidata.org/wiki/Q1668499","display_name":"Rewriting","level":2,"score":0.2865999937057495},{"id":"https://openalex.org/C133462117","wikidata":"https://www.wikidata.org/wiki/Q4929239","display_name":"Data collection","level":2,"score":0.26260000467300415},{"id":"https://openalex.org/C3018395757","wikidata":"https://www.wikidata.org/wiki/Q1379672","display_name":"Evaluation methods","level":2,"score":0.2614000141620636},{"id":"https://openalex.org/C115901376","wikidata":"https://www.wikidata.org/wiki/Q184199","display_name":"Automation","level":2,"score":0.2612999975681305},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.2612999975681305},{"id":"https://openalex.org/C2987098735","wikidata":"https://www.wikidata.org/wiki/Q3808900","display_name":"Recall rate","level":2,"score":0.2605000138282776},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.25690001249313354},{"id":"https://openalex.org/C199521495","wikidata":"https://www.wikidata.org/wiki/Q181487","display_name":"Audit","level":2,"score":0.2522999942302704}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.24839","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.24839","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.24839","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.24839","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":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":{"Agentic":[0],"data":[1],"analysis":[2],"systems":[3],"produce":[4],"rich":[5],"outputs,":[6],"including":[7],"code,":[8],"numerical":[9,69],"results,":[10],"and":[11,37,54,76,89,96,157,167,207],"verbal":[12],"diagnostics.":[13],"This":[14],"makes":[15],"them":[16],"more":[17],"challenging":[18],"to":[19,29,155,162],"evaluate":[20,77],"than":[21],"single-turn":[22],"LLM":[23],"responses.":[24],"It":[25],"is":[26,117,142,195],"therefore":[27],"necessary":[28],"distinguish":[30],"genuine":[31],"disagreement":[32],"between":[33],"an":[34,181],"agent's":[35],"output":[36],"a":[38,52,63,78,136],"ground-truth":[39],"answer":[40,182],"from":[41,72,153,160],"grading":[42,58,81,150,204],"artifacts.":[43],"We":[44,74,185],"investigate":[45],"how":[46],"reliably":[47],"automated":[48,104],"graders":[49,105],"assess":[50],"such":[51],"system":[53],"what":[55],"strategies":[56,98],"improve":[57],"quality":[59],"by":[60,131],"applying":[61],"LAMBDA,":[62],"multi-agent":[64],"data-analysis":[65],"system,":[66],"on":[67],"153":[68],"QRData":[70],"tasks":[71],"DSGym.":[73],"develop":[75],"three-layer":[79],"human-AI":[80],"cascade:":[82],"strict":[83,128],"regex":[84],"matching,":[85],"LLM-based":[86],"lenient":[87,114,140],"grading,":[88],"snippet-based":[90],"human":[91,120],"inspection,":[92],"which":[93],"combines":[94],"non-GenAI":[95],"GenAI":[97],"with":[99,166,203],"different":[100],"failure":[101],"profiles.":[102],"Both":[103],"achieve":[106],"100%":[107],"observed":[108,208],"precision":[109],"(0/70":[110],"false":[111],"positives).":[112],"The":[113],"grader's":[115,129],"recall":[116,130],"97%":[118,156],"against":[119],"labels.":[121],"A":[122],"keyword-anchored":[123],"extraction":[124],"pipeline":[125,205],"raises":[126,149],"the":[127,139,178,196],"60":[132],"percentage":[133],"points":[134],"over":[135],"last-number":[137],"heuristic;":[138],"grader":[141],"architecturally":[143],"parser-independent.":[144],"An":[145],"iterative":[146],"nudge":[147,179],"mechanism":[148],"run":[151],"success":[152],"36%":[154],"lenient-pass":[158],"rates":[159],"16%":[161],"46%;":[163],"comparing":[164],"nudging":[165],"without":[168],"original-question":[169],"re-injection":[170,173],"shows":[171],"that":[172,192],"offers":[174],"no":[175],"benefit,":[176],"confirming":[177],"as":[180],"template":[183],"cue.":[184],"further":[186],"observe":[187],"in":[188],"this":[189],"case":[190],"study":[191],"variable":[193],"type":[194],"task":[197],"metadata":[198],"field":[199],"most":[200],"consistently":[201],"associated":[202],"dynamics":[206],"outcome":[209],"grades.":[210]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-25T00:00:00"}