{"id":"https://openalex.org/W7139916360","doi":"https://doi.org/10.48550/arxiv.2603.18662","title":"Thinking with Constructions: A Benchmark and Policy Optimization for Visual-Text Interleaved Geometric Reasoning","display_name":"Thinking with Constructions: A Benchmark and Policy Optimization for Visual-Text Interleaved Geometric Reasoning","publication_year":2026,"publication_date":"2026-03-19","ids":{"openalex":"https://openalex.org/W7139916360","doi":"https://doi.org/10.48550/arxiv.2603.18662"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.18662","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.18662","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":null,"license_id":null,"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.2603.18662","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5111174926","display_name":"Haokun Zhao","orcid":"https://orcid.org/0009-0006-5069-8840"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Haokun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130234579","display_name":"Wanshi Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Wanshi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130240258","display_name":"Haidong Yuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Haidong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050836381","display_name":"Songjun Cao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cao, Songjun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130229453","display_name":"Long Ma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Long","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5130223266","display_name":"Yanghua Xiao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xiao, Yanghua","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":0.5598000288009644,"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":0.5598000288009644,"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/T10799","display_name":"Data Visualization and Analytics","score":0.10350000113248825,"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/T11596","display_name":"Constraint Satisfaction and Optimization","score":0.04769999906420708,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/leverage","display_name":"Leverage (statistics)","score":0.7177000045776367},{"id":"https://openalex.org/keywords/counterfactual-thinking","display_name":"Counterfactual thinking","score":0.7172999978065491},{"id":"https://openalex.org/keywords/visual-reasoning","display_name":"Visual reasoning","score":0.680400013923645},{"id":"https://openalex.org/keywords/benchmark","display_name":"Benchmark (surveying)","score":0.5555999875068665},{"id":"https://openalex.org/keywords/entropy","display_name":"Entropy (arrow of time)","score":0.529699981212616},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5148000121116638},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5088000297546387},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.4117000102996826},{"id":"https://openalex.org/keywords/coding","display_name":"Coding (social sciences)","score":0.3912999927997589}],"concepts":[{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.7177000045776367},{"id":"https://openalex.org/C108650721","wikidata":"https://www.wikidata.org/wiki/Q1783253","display_name":"Counterfactual thinking","level":2,"score":0.7172999978065491},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.685699999332428},{"id":"https://openalex.org/C2777508537","wikidata":"https://www.wikidata.org/wiki/Q7936620","display_name":"Visual reasoning","level":2,"score":0.680400013923645},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.5555999875068665},{"id":"https://openalex.org/C106301342","wikidata":"https://www.wikidata.org/wiki/Q4117933","display_name":"Entropy (arrow of time)","level":2,"score":0.529699981212616},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5289999842643738},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5148000121116638},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5088000297546387},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.4117000102996826},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3968999981880188},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.3912999927997589},{"id":"https://openalex.org/C94966114","wikidata":"https://www.wikidata.org/wiki/Q29256","display_name":"Black box","level":2,"score":0.3206999897956848},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.30140000581741333},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3003000020980835},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.2994000017642975},{"id":"https://openalex.org/C166088908","wikidata":"https://www.wikidata.org/wiki/Q308495","display_name":"Abductive reasoning","level":2,"score":0.2978000044822693},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.29019999504089355},{"id":"https://openalex.org/C2777904410","wikidata":"https://www.wikidata.org/wiki/Q7397","display_name":"Software","level":2,"score":0.28700000047683716},{"id":"https://openalex.org/C100776233","wikidata":"https://www.wikidata.org/wiki/Q2532492","display_name":"Bridge (graph theory)","level":2,"score":0.28619998693466187},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.28110000491142273},{"id":"https://openalex.org/C2776224158","wikidata":"https://www.wikidata.org/wiki/Q187931","display_name":"Phrase","level":2,"score":0.2809000015258789},{"id":"https://openalex.org/C90509273","wikidata":"https://www.wikidata.org/wiki/Q11012","display_name":"Robot","level":2,"score":0.28029999136924744},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.27250000834465027},{"id":"https://openalex.org/C4554734","wikidata":"https://www.wikidata.org/wiki/Q593744","display_name":"Knowledge base","level":2,"score":0.26829999685287476},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.26420000195503235},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.26010000705718994},{"id":"https://openalex.org/C56461940","wikidata":"https://www.wikidata.org/wiki/Q970687","display_name":"Eye tracking","level":2,"score":0.2524000108242035}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.18662","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.18662","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.18662","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.18662","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":null,"license_id":null,"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":{"Geometric":[0],"reasoning":[1,115],"inherently":[2],"requires":[3],"\"thinking":[4],"with":[5,36,79,113],"constructions\"":[6],"--":[7],"the":[8,16,40,67,143],"dynamic":[9],"manipulation":[10],"of":[11,43,147],"visual":[12,50,81,148],"aids":[13,93,149],"to":[14,33,47,141,153,165],"bridge":[15],"gap":[17],"between":[18],"problem":[19],"conditions":[20],"and":[21,45,103,145,178],"solutions.":[22],"However,":[23],"existing":[24],"Multimodal":[25],"Large":[26],"Language":[27],"Models":[28],"(MLLMs)":[29],"are":[30,180],"largely":[31],"confined":[32],"passive":[34],"inference":[35],"static":[37],"diagrams,":[38],"lacking":[39],"strategic":[41,134],"knowledge":[42],"when":[44],"how":[46],"construct":[48],"effective":[49],"aids.":[51],"To":[52],"address":[53],"this,":[54],"we":[55,121],"present":[56],"a":[57,128,171],"framework":[58],"for":[59,132],"Visual-Text":[60],"Interleaved":[61],"Chain-of-Thought.":[62],"We":[63],"first":[64,68],"introduce":[65],"GeoAux-Bench,":[66],"benchmark":[69],"comprising":[70],"4,334":[71],"geometry":[72],"problems":[73],"that":[74],"aligns":[75],"textual":[76],"construction":[77],"steps":[78],"ground-truth":[80],"updates.":[82],"Our":[83],"pilot":[84],"study":[85],"reveals":[86],"two":[87],"critical":[88],"insights:":[89],"(1)":[90],"interleaved":[91],"visual-textual":[92],"outperform":[94],"single-modality":[95],"counterparts,":[96],"which":[97],"cannot":[98],"losslessly":[99],"capture":[100],"geometric":[101],"synergy;":[102],"(2)":[104],"valid":[105],"constructions":[106],"act":[107],"as":[108],"entropy":[109],"reducers,":[110],"strongly":[111],"correlating":[112],"reduced":[114],"perplexity.":[116],"Building":[117],"on":[118,182],"these":[119],"findings,":[120],"propose":[122],"Action":[123],"Applicability":[124],"Policy":[125],"Optimization":[126],"(A2PO),":[127],"reinforcement":[129],"learning":[130],"paradigm":[131],"mastering":[133],"construction.":[135],"A2PO":[136],"employs":[137],"Adaptive":[138],"Reward":[139],"Shaping":[140],"regulate":[142],"timing":[144],"quality":[146],"via":[150],"counterfactual":[151],"sampling":[152],"distinguish":[154],"necessary":[155],"from":[156],"redundant":[157],"constructions.":[158],"Experiments":[159],"demonstrate":[160],"our":[161],"approach":[162],"enables":[163],"MLLMs":[164],"leverage":[166],"selective":[167],"auxiliary":[168],"constructions,":[169],"yielding":[170],"3.51%":[172],"gain":[173],"over":[174],"strong":[175],"baselines.":[176],"Code":[177],"data":[179],"available":[181],"GitHub.":[183]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-21T00:00:00"}