{"id":"https://openalex.org/W7225150694","doi":"https://doi.org/10.48550/arxiv.2611.03097","title":"DR Pixel: Page-Native Deep Research with Vision-Language Models","display_name":"DR Pixel: Page-Native Deep Research with Vision-Language Models","publication_year":2026,"publication_date":"2026-11-02","ids":{"openalex":"https://openalex.org/W7225150694","doi":"https://doi.org/10.48550/arxiv.2611.03097"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2611.03097","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2611.03097","pdf_url":"https://arxiv.org/pdf/2611.03097","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2611.03097","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5119963925","display_name":"Tiansheng Hu","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hu, Tiansheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160445395","display_name":"Yichuan Wang","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Yichuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5029735169","display_name":"James V. Talwar","orcid":"https://orcid.org/0000-0002-9520-1950","observed_orcids":["https://orcid.org/0000-0002-9520-1950"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Talwar, James V.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013447238","display_name":"Nicholas Tomlin","orcid":"https://orcid.org/0000-0001-6502-989X","observed_orcids":["https://orcid.org/0000-0001-6502-989X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tomlin, Nicholas","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5160466980","display_name":"Chen Zhao","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5039158419","display_name":"Sewon Min","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Min, Sewon","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.8529999852180481,"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":0.8529999852180481,"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/T10286","display_name":"Information Retrieval and Search Behavior","score":0.06449999660253525,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T10028","display_name":"Topic Modeling","score":0.021299999207258224,"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/deep-research-agents","display_name":"deep research agents","score":0.7900000214576721},{"id":"https://openalex.org/keywords/vision-language-models","display_name":"vision-language models","score":0.9430000185966492},{"id":"https://openalex.org/keywords/scientific-publications","display_name":"scientific publications","score":0.7509999871253967},{"id":"https://openalex.org/keywords/multimodal-agents","display_name":"multimodal agents","score":0.6259999871253967},{"id":"https://openalex.org/keywords/teacher-trajectories","display_name":"teacher trajectories","score":0.4830000102519989},{"id":"https://openalex.org/keywords/document-images","display_name":"scanned documents","score":0.4410000145435333},{"id":"https://openalex.org/keywords/tables-and-figures","display_name":"tables and figures","score":0.5389999747276306},{"id":"https://openalex.org/keywords/equations","display_name":"equations","score":0.5740000009536743}],"concepts":[{"id":"https://openalex.org/C108650721","wikidata":"https://www.wikidata.org/wiki/Q1783253","display_name":"Counterfactual thinking","level":2,"score":0.7348999977111816},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7077999711036682},{"id":"https://openalex.org/C55166926","wikidata":"https://www.wikidata.org/wiki/Q2892946","display_name":"Oracle","level":2,"score":0.6424999833106995},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5864999890327454},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5683000087738037},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5659999847412109},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.544700026512146},{"id":"https://openalex.org/C100660578","wikidata":"https://www.wikidata.org/wiki/Q18733","display_name":"Recall","level":2,"score":0.46810001134872437},{"id":"https://openalex.org/C81669768","wikidata":"https://www.wikidata.org/wiki/Q2359161","display_name":"Precision and recall","level":2,"score":0.45089998841285706},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.4487000107765198},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4480000138282776},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.4375999867916107},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.41530001163482666},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.36570000648498535},{"id":"https://openalex.org/C44291984","wikidata":"https://www.wikidata.org/wiki/Q1074173","display_name":"Question answering","level":2,"score":0.3264999985694885},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.31040000915527344},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.31029999256134033},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.30660000443458557},{"id":"https://openalex.org/C98184364","wikidata":"https://www.wikidata.org/wiki/Q1780131","display_name":"Argument (complex analysis)","level":2,"score":0.2831999957561493},{"id":"https://openalex.org/C2992562121","wikidata":"https://www.wikidata.org/wiki/Q3817808","display_name":"Scientific reasoning","level":2,"score":0.27549999952316284},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.2669000029563904},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2635999917984009},{"id":"https://openalex.org/C2777206241","wikidata":"https://www.wikidata.org/wiki/Q194431","display_name":"Paragraph","level":2,"score":0.2632000148296356},{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.25780001282691956},{"id":"https://openalex.org/C2777530160","wikidata":"https://www.wikidata.org/wiki/Q41796","display_name":"Sentence","level":2,"score":0.2540000081062317},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.25189998745918274}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2611.03097","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2611.03097","pdf_url":"https://arxiv.org/pdf/2611.03097","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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.2611.03097","doi":"https://doi.org/10.48550/arxiv.2611.03097","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2611.03097","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2611.03097","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2611.03097","pdf_url":"https://arxiv.org/pdf/2611.03097","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4400373146","https://openalex.org/W7197019678","https://openalex.org/W7133243847","https://openalex.org/W7166706851","https://openalex.org/W4416785301","https://openalex.org/W4378468513","https://openalex.org/W7222553824","https://openalex.org/W7137959380","https://openalex.org/W4414810701","https://openalex.org/W4416036851"],"abstract_inverted_index":{"Complex":[0],"documents":[1],"such":[2],"as":[3,63],"scientific":[4,87],"papers":[5,22,88],"contain":[6],"rich":[7],"information":[8],"in":[9],"tables,":[10,94],"figures,":[11,93],"and":[12,40,67,95,114,143,168,188,197,242,254],"equations;":[13],"however,":[14],"existing":[15,205],"deep":[16,36],"research":[17],"(DR)":[18],"agents":[19],"typically":[20],"reduce":[21],"to":[23,124,175],"OCR-extracted":[24],"text.":[25],"We":[26,47,121],"instead":[27],"ask":[28],"whether":[29],"vision-language":[30],"models":[31,110],"(VLMs)":[32],"can":[33],"enable":[34],"page-native":[35,98],"research:":[37],"directly":[38,240],"retrieving":[39,241],"reasoning":[41,68,90,142,167,243],"over":[42,71,91,192,244],"rendered":[43,60],"document":[44,246],"page":[45,61],"images.":[46],"introduce":[48,155],"DR":[49,56,99,125,210,250],"Pixel,":[50],"the":[51,176,200,222,230],"first":[52],"open-source":[53],"end-to-end":[54],"multimodal":[55],"agent":[57],"that":[58,131,164,219,229],"uses":[59],"images":[62],"its":[64],"sole":[65],"retrieval":[66,182],"unit,":[69],"operating":[70],"a":[72,84,159],"corpus":[73],"of":[74,216],"1.06M":[75],"pages.":[76],"To":[77,151],"evaluate":[78],"this":[79],"setting,":[80],"we":[81,154],"construct":[82],"DRPixelBench,":[83,196],"benchmark":[85],"on":[86,195,203],"requiring":[89,117],"text,":[92],"equations.":[96],"Surprisingly,":[97],"already":[100],"works":[101],"well":[102],"without":[103],"training,":[104],"outperforming":[105],"OCR-based":[106],"baselines":[107],"across":[108],"state-of-the-art":[109],"like":[111],"GPT-5.4,":[112],"Gemini-3.5-Flash,":[113],"Qwen3.6-27B,":[115],"while":[116],"fewer":[118],"tool":[119],"calls.":[120],"next":[122],"turn":[123],"training.":[126],"A":[127],"key":[128],"challenge":[129],"is":[130,220,234],"teacher":[132],"trajectories":[133],"for":[134],"agentic":[135],"supervised":[136],"finetuning":[137],"are":[138],"noisy,":[139],"mixing":[140],"critical":[141],"action":[144,169],"steps":[145,170],"with":[146],"redundant":[147],"or":[148],"failed":[149],"attempts.":[150],"address":[152],"this,":[153],"counterfactual":[156,173],"credit":[157,161],"reweighting,":[158],"fine-grained":[160],"assignment":[162],"method":[163],"weights":[165],"individual":[166],"by":[171,184,191],"their":[172],"contribution":[174],"final":[177,189],"answer.":[178],"Our":[179,226],"training":[180],"improves":[181],"recall":[183],"about":[185],"15":[186],"points":[187,194,215],"accuracy":[190],"10":[193],"likewise":[198],"outperforms":[199],"untrained":[201],"baseline":[202],"four":[204],"benchmarks.":[206],"On":[207],"out-of-distribution":[208],"ViDoRe-V3,":[209],"Pixel":[211],"comes":[212],"within":[213],"1.8":[214],"an":[217],"oracle":[218],"given":[221],"gold":[223],"pages":[224,247],"directly.":[225],"results":[227],"suggest":[228],"conventional":[231],"OCR-then-retrieve":[232],"pipeline":[233],"not":[235],"only":[236],"unnecessary":[237],"but":[238],"suboptimal:":[239],"visual":[245],"makes":[248],"complex":[249],"both":[251],"more":[252,255],"effective":[253],"efficient.":[256]},"counts_by_year":[],"updated_date":"2026-11-05T12:25:35.769452","created_date":"2026-11-04T00:00:00"}