{"id":"https://openalex.org/W7131874110","doi":"https://doi.org/10.48550/arxiv.2602.22703","title":"Enhancing Geometric Perception in VLMs via Translator-Guided Reinforcement Learning","display_name":"Enhancing Geometric Perception in VLMs via Translator-Guided Reinforcement Learning","publication_year":2026,"publication_date":"2026-02-26","ids":{"openalex":"https://openalex.org/W7131874110","doi":"https://doi.org/10.48550/arxiv.2602.22703"},"language":null,"primary_location":{"id":"pmh:doi:10.48550/arxiv.2602.22703","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":null,"any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5127308263","display_name":"Hao Yu","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Yu, Hao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127376955","display_name":"Shuning Jia","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jia, Shuning","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127297016","display_name":"Guanghao Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Guanghao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127103462","display_name":"Wenhao Jiang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jiang, Wenhao","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5127161534","display_name":"Chun Yuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Chun","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5127308263"],"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":0.7997000217437744,"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":0.7997000217437744,"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/T10028","display_name":"Topic Modeling","score":0.04149999842047691,"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/T10653","display_name":"Robot Manipulation and Learning","score":0.018200000748038292,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.7103000283241272},{"id":"https://openalex.org/keywords/perception","display_name":"Perception","score":0.6344000101089478},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.6089000105857849},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.4498000144958496},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4284000098705292},{"id":"https://openalex.org/keywords/bridge","display_name":"Bridge (graph theory)","score":0.387800008058548},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.38760000467300415},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.37689998745918274},{"id":"https://openalex.org/keywords/visual-perception","display_name":"Visual perception","score":0.3693999946117401}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7103000283241272},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6700999736785889},{"id":"https://openalex.org/C26760741","wikidata":"https://www.wikidata.org/wiki/Q160402","display_name":"Perception","level":2,"score":0.6344000101089478},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.6089000105857849},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5996999740600586},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4731000065803528},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.4498000144958496},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4284000098705292},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.387800008058548},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.38760000467300415},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.37689998745918274},{"id":"https://openalex.org/C178253425","wikidata":"https://www.wikidata.org/wiki/Q162668","display_name":"Visual perception","level":3,"score":0.3693999946117401},{"id":"https://openalex.org/C186399060","wikidata":"https://www.wikidata.org/wiki/Q959962","display_name":"Diagram","level":2,"score":0.3686999976634979},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3278000056743622},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.31439998745918274},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.30959999561309814},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.2992999851703644},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.298799991607666},{"id":"https://openalex.org/C104065381","wikidata":"https://www.wikidata.org/wiki/Q1002535","display_name":"Geometric modeling","level":2,"score":0.2816999852657318},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.2793999910354614},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.27639999985694885},{"id":"https://openalex.org/C136389625","wikidata":"https://www.wikidata.org/wiki/Q334384","display_name":"Supervised learning","level":3,"score":0.275299996137619},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.27309998869895935},{"id":"https://openalex.org/C136520226","wikidata":"https://www.wikidata.org/wiki/Q302814","display_name":"Geometric data analysis","level":2,"score":0.26190000772476196},{"id":"https://openalex.org/C7305733","wikidata":"https://www.wikidata.org/wiki/Q207961","display_name":"Geometric shape","level":2,"score":0.26179999113082886},{"id":"https://openalex.org/C22033958","wikidata":"https://www.wikidata.org/wiki/Q7167036","display_name":"Perceptual learning","level":3,"score":0.2574000060558319},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2533999979496002}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:doi:10.48550/arxiv.2602.22703","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},{"id":"doi:10.48550/arxiv.2602.22703","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.22703","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":"article"}],"best_oa_location":{"id":"pmh:doi:10.48550/arxiv.2602.22703","is_oa":true,"landing_page_url":null,"pdf_url":null,"source":{"id":"https://openalex.org/S4406922384","display_name":"Open MIND","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":null,"host_organization_name":null,"host_organization_lineage":[],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Article"},"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":{"Vision-language":[0],"models":[1],"(VLMs)":[2],"often":[3],"struggle":[4],"with":[5,30,36],"geometric":[6,50,65,132],"reasoning":[7,175],"due":[8],"to":[9,97,196],"their":[10],"limited":[11],"perception":[12,51,66,133],"of":[13,49,68,95,108,186],"fundamental":[14],"diagram":[15,27],"elements.":[16],"To":[17,55],"tackle":[18],"this":[19],"challenge,":[20],"we":[21,70],"introduce":[22],"GeoPerceive,":[23,96],"a":[24,73],"benchmark":[25],"comprising":[26],"instances":[28],"paired":[29],"domain-specific":[31],"language":[32,100],"(DSL)":[33],"representations,":[34],"along":[35],"an":[37,81],"efficient":[38],"automatic":[39],"data":[40,58,93],"generation":[41],"pipeline.":[42],"This":[43,103],"design":[44],"enables":[45],"the":[46,57,64,92,106,180],"isolated":[47],"evaluation":[48],"independently":[52],"from":[53],"reasoning.":[54,138],"exploit":[56],"provided":[59],"by":[60,91],"GeoPerceive":[61],"for":[62],"enhancing":[63],"capabilities":[67],"VLMs,":[69],"propose":[71],"GeoDPO,":[72],"translator-guided":[74],"reinforcement":[75,118],"learning":[76],"(RL)":[77],"framework.":[78],"GeoDPO":[79,122,159,187],"employs":[80],"NL-to-DSL":[82],"translator,":[83],"which":[84,112],"is":[85],"trained":[86],"on":[87,123,164,168,173],"synthetic":[88],"pairs":[89],"generated":[90],"engine":[94],"bridge":[98],"natural":[99],"and":[101,126,151,171,183],"DSL.":[102],"translator":[104],"facilitates":[105],"computation":[107],"fine-grained,":[109],"DSL-level":[110],"scores,":[111],"serve":[113],"as":[114,134,136],"reward":[115],"signals":[116],"in":[117,131,156],"learning.":[119],"We":[120],"assess":[121],"both":[124],"in-domain":[125,165],"out-of-domain":[127,157,169],"datasets,":[128],"spanning":[129],"tasks":[130],"well":[135],"downstream":[137,174],"Experimental":[139],"results":[140],"demonstrate":[141],"that,":[142],"while":[143],"supervised":[144],"fine-tuning":[145],"(SFT)":[146],"offers":[147],"only":[148],"marginal":[149],"improvements":[150],"may":[152],"even":[153],"impair":[154],"performance":[155,182],"scenarios,":[158],"achieves":[160],"substantial":[161],"gains:":[162],"$+26.5\\%$":[163],"data,":[166,170],"$+8.0\\%$":[167],"$+39.0\\%$":[172],"tasks.":[176],"These":[177],"findings":[178],"underscore":[179],"superior":[181],"generalization":[184],"ability":[185],"over":[188],"SFT.":[189],"All":[190],"codes":[191],"are":[192],"released":[193],"at":[194],"https://github.com/Longin-Yu/GeoPerceive":[195],"ensure":[197],"reproducibility.":[198]},"counts_by_year":[],"updated_date":"2026-04-04T16:13:02.066488","created_date":"2026-02-28T00:00:00"}