{"id":"https://openalex.org/W7139097793","doi":"https://doi.org/10.48550/arxiv.2603.16189","title":"Reliable Reasoning in SVG-LLMs via Multi-Task Multi-Reward Reinforcement Learning","display_name":"Reliable Reasoning in SVG-LLMs via Multi-Task Multi-Reward Reinforcement Learning","publication_year":2026,"publication_date":"2026-03-17","ids":{"openalex":"https://openalex.org/W7139097793","doi":"https://doi.org/10.48550/arxiv.2603.16189"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.16189","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.16189","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.2603.16189","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129815534","display_name":"Haomin Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Haomin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130059803","display_name":"Qi Wei","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wei, Qi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5129843047","display_name":"Qianli Ma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Qianli","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130148236","display_name":"Shengyuan Ding","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ding, Shengyuan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102216495","display_name":"Jinhui Yin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Jinhui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5130127028","display_name":"Kai Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Kai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5129879464","display_name":"Hongjie Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Hongjie","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":0.5848000049591064,"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":0.5848000049591064,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.07320000231266022,"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/T12262","display_name":"Hate Speech and Cyberbullying Detection","score":0.03790000081062317,"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/scalable-vector-graphics","display_name":"Scalable Vector Graphics","score":0.8651000261306763},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.6245999932289124},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.5799000263214111},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.4747999906539917},{"id":"https://openalex.org/keywords/construct","display_name":"Construct (python library)","score":0.47269999980926514},{"id":"https://openalex.org/keywords/coherence","display_name":"Coherence (philosophical gambling strategy)","score":0.44350001215934753},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4375}],"concepts":[{"id":"https://openalex.org/C202629362","wikidata":"https://www.wikidata.org/wiki/Q2078","display_name":"Scalable Vector Graphics","level":2,"score":0.8651000261306763},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7958999872207642},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.6245999932289124},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5799000263214111},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5033000111579895},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4909000098705292},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.4747999906539917},{"id":"https://openalex.org/C2780801425","wikidata":"https://www.wikidata.org/wiki/Q5164392","display_name":"Construct (python library)","level":2,"score":0.47269999980926514},{"id":"https://openalex.org/C2781181686","wikidata":"https://www.wikidata.org/wiki/Q4226068","display_name":"Coherence (philosophical gambling strategy)","level":2,"score":0.44350001215934753},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4375},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.374099999666214},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.3434000015258789},{"id":"https://openalex.org/C133162039","wikidata":"https://www.wikidata.org/wiki/Q1061077","display_name":"Code generation","level":3,"score":0.33309999108314514},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.33070001006126404},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.2838999927043915},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.27559998631477356},{"id":"https://openalex.org/C18555067","wikidata":"https://www.wikidata.org/wiki/Q8375051","display_name":"Joint (building)","level":2,"score":0.2734000086784363},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2653999924659729},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2597000002861023}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.16189","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.16189","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.2603.16189","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.16189","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":{"With":[0],"the":[1,73,104,123],"rapid":[2],"advancement":[3],"of":[4,10,48],"vision-language":[5],"models,":[6],"an":[7],"increasing":[8],"number":[9],"studies":[11],"have":[12],"explored":[13],"their":[14],"potential":[15],"for":[16,60],"SVG":[17,28,78,95,110,169],"generation":[18,153],"tasks.":[19,101],"Although":[20],"existing":[21,161],"approaches":[22],"improve":[23],"performance":[24],"by":[25],"constructing":[26],"large-scale":[27],"datasets":[29],"and":[30,45,99,117,126,137,145,172],"introducing":[31],"SVG-specific":[32],"tokens,":[33],"they":[34],"still":[35],"suffer":[36],"from":[37],"limited":[38],"generalization,":[39],"redundant":[40],"paths":[41],"in":[42],"code":[43,96,138,170],"outputs,":[44],"a":[46,62,67,88,128],"lack":[47],"explicit":[49],"reasoning.":[50],"In":[51],"this":[52,82],"work,":[53],"we":[54,85,121],"present":[55],"CTRL-S":[56,112,159],"(Chain-of-Thought":[57],"Reinforcement":[58],"Learning":[59],"SVG),":[61],"unified":[63],"framework":[64],"that":[65,158],"introduces":[66],"chain-of-thought":[68],"mechanism":[69],"to":[70,106],"explicitly":[71],"expose":[72],"model's":[74],"reasoning":[75],"process":[76],"during":[77],"generation.":[79],"To":[80],"support":[81],"structured":[83,109],"reasoning,":[84],"construct":[86],"SVG-Sophia,":[87],"high-quality":[89],"dataset":[90],"containing":[91],"145K":[92],"samples":[93],"across":[94],"refinement,":[97],"Text-to-SVG,":[98],"Image-to-SVG":[100],"By":[102],"training":[103],"model":[105],"generate":[107],"group-level":[108],"code,":[111],"significantly":[113],"improves":[114],"structural":[115],"coherence":[116],"visual":[118,174],"fidelity.":[119,175],"Furthermore,":[120],"adopt":[122],"GRPO":[124],"algorithm":[125],"design":[127],"multi-reward":[129,143],"optimization":[130,144],"framework,":[131],"incorporating":[132],"DINO,":[133],"image-text":[134],"similarity,":[135],"format,":[136],"efficiency":[139],"rewards.":[140],"Through":[141],"joint":[142],"multi-task":[146],"training,":[147],"our":[148],"approach":[149],"systematically":[150],"enhances":[151],"overall":[152],"capabilities.":[154],"Extensive":[155],"experiments":[156],"show":[157],"outperforms":[160],"methods,":[162],"achieving":[163],"higher":[164],"task":[165],"success":[166],"rates,":[167],"superior":[168],"quality,":[171],"exceptional":[173]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-03-20T00:00:00"}