{"id":"https://openalex.org/W7167098711","doi":"https://doi.org/10.48550/arxiv.2607.00924","title":"Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination","display_name":"Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination","publication_year":2026,"publication_date":"2026-07-01","ids":{"openalex":"https://openalex.org/W7167098711","doi":"https://doi.org/10.48550/arxiv.2607.00924"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.00924","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00924","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,"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.2607.00924","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139925359","display_name":"Subhadeep Pal","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pal, Subhadeep","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139934297","display_name":"Shashwat Sourav","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sourav, Shashwat","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139868173","display_name":"Tirthankar Ghosal","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ghosal, Tirthankar","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5011504360","display_name":"Markus J. Buehler","orcid":"https://orcid.org/0000-0002-4173-9659"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Buehler, Markus J.","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/T11948","display_name":"Machine Learning in Materials Science","score":0.9031000137329102,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T11948","display_name":"Machine Learning in Materials Science","score":0.9031000137329102,"subfield":{"id":"https://openalex.org/subfields/2505","display_name":"Materials Chemistry"},"field":{"id":"https://openalex.org/fields/25","display_name":"Materials Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11273","display_name":"Advanced Graph Neural Networks","score":0.061900001019239426,"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/T10028","display_name":"Topic Modeling","score":0.00430000014603138,"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/reinforcement-learning","display_name":"Reinforcement learning","score":0.5389000177383423},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.445499986410141},{"id":"https://openalex.org/keywords/graph","display_name":"Graph","score":0.40959998965263367},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.3716000020503998},{"id":"https://openalex.org/keywords/knowledge-base","display_name":"Knowledge base","score":0.3619999885559082},{"id":"https://openalex.org/keywords/semantic-memory","display_name":"Semantic memory","score":0.35670000314712524},{"id":"https://openalex.org/keywords/mechanism","display_name":"Mechanism (biology)","score":0.34850001335144043}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6470999717712402},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5422000288963318},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5389000177383423},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.445499986410141},{"id":"https://openalex.org/C132525143","wikidata":"https://www.wikidata.org/wiki/Q141488","display_name":"Graph","level":2,"score":0.40959998965263367},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.3716000020503998},{"id":"https://openalex.org/C4554734","wikidata":"https://www.wikidata.org/wiki/Q593744","display_name":"Knowledge base","level":2,"score":0.3619999885559082},{"id":"https://openalex.org/C197914299","wikidata":"https://www.wikidata.org/wiki/Q18650","display_name":"Semantic memory","level":3,"score":0.35670000314712524},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3553999960422516},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.34850001335144043},{"id":"https://openalex.org/C234837","wikidata":"https://www.wikidata.org/wiki/Q1420493","display_name":"Conceptual graph","level":3,"score":0.3328000009059906},{"id":"https://openalex.org/C198942812","wikidata":"https://www.wikidata.org/wiki/Q496618","display_name":"Semantic property","level":2,"score":0.3319000005722046},{"id":"https://openalex.org/C2129575","wikidata":"https://www.wikidata.org/wiki/Q54837","display_name":"Semantic Web","level":2,"score":0.295199990272522},{"id":"https://openalex.org/C188147891","wikidata":"https://www.wikidata.org/wiki/Q147638","display_name":"Cognitive science","level":1,"score":0.2847000062465668},{"id":"https://openalex.org/C67277372","wikidata":"https://www.wikidata.org/wiki/Q7449085","display_name":"Semantic role labeling","level":3,"score":0.2842999994754791},{"id":"https://openalex.org/C85407183","wikidata":"https://www.wikidata.org/wiki/Q1045785","display_name":"Semantic network","level":2,"score":0.2728999853134155},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.26660001277923584},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2651999890804291}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.00924","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00924","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,"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.2607.00924","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00924","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,"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":[{"id":"https://metadata.un.org/sdg/4","score":0.7524584531784058,"display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Accelerating":[0],"materials":[1,29,102,195],"discovery":[2],"requires":[3],"AI":[4,188],"systems":[5,189],"that":[6,158],"can":[7],"generate":[8],"scientifically":[9],"valid":[10],"hypotheses":[11],"through":[12],"multi-step,":[13],"domain-grounded":[14],"reasoning.":[15,45],"Standard":[16],"large":[17],"language":[18,82],"models":[19,54],"often":[20],"produce":[21],"fluent":[22],"but":[23],"weakly":[24],"traceable":[25],"responses":[26],"to":[27,35,62,91],"open-ended":[28,99],"design":[30,79,196],"problems,":[31],"making":[32],"it":[33],"difficult":[34],"determine":[36],"whether":[37],"final":[38,151],"answers":[39],"are":[40],"supported":[41],"by":[42],"coherent":[43],"intermediate":[44],"We":[46],"develop":[47],"Graph-PRefLexOR,":[48],"a":[49,167,184],"family":[50],"of":[51],"graph-native":[52,180],"reasoning":[53,64,120,149],"fine-tuned":[55],"with":[56,84,115],"Group":[57],"Relative":[58],"Policy":[59],"Optimization":[60],"(GRPO)":[61],"organize":[63],"into":[65],"explicit":[66],"phases":[67],"for":[68,190],"mechanism":[69],"exploration,":[70],"graph":[71,155],"construction,":[72],"pattern":[73],"extraction,":[74],"and":[75,95,104,128,139,150,197],"hypothesis":[76,192],"synthesis.":[77],"This":[78],"links":[80],"neural":[81],"generation":[83,193],"symbolic":[85],"relational":[86],"structure,":[87],"enabling":[88],"causal":[89],"connections":[90],"be":[92],"constructed,":[93],"inspected,":[94],"reused.":[96],"On":[97],"100":[98],"questions":[100],"from":[101],"science":[103],"mechanics":[105],"literature,":[106],"Graph-PRefLexOR":[107],"achieves":[108],"40-65%":[109],"improvements":[110],"over":[111],"corresponding":[112],"base":[113],"models,":[114],"the":[116],"largest":[117],"gains":[118],"in":[119,194],"traceability.":[121],"Embedding":[122],"analyses":[123,142],"show":[124,144],"broader":[125],"semantic":[126,133,169,175],"exploration":[127],"approximately":[129],"2-3":[130],"times":[131],"greater":[132],"diversity":[134],"than":[135,172],"baselines.":[136],"Semantic":[137],"backtracking":[138],"layer-wise":[140],"hidden-state":[141],"further":[143],"stronger":[145],"alignment":[146],"between":[147],"structured":[148],"answers.":[152],"Finally,":[153],"test-time":[154],"expansion":[156],"reveals":[157],"additional":[159],"compute":[160],"primarily":[161],"increases":[162],"long-range":[163],"conceptual":[164],"recombination":[165],"within":[166],"bounded":[168],"space,":[170],"rather":[171],"simply":[173],"expanding":[174],"coverage.":[176],"These":[177],"results":[178],"establish":[179],"reinforcement":[181],"learning":[182],"as":[183],"pathway":[185],"toward":[186],"interpretable":[187],"scientific":[191,199],"other":[198],"applications.":[200]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-03T00:00:00"}