{"id":"https://openalex.org/W7167032472","doi":"https://doi.org/10.48550/arxiv.2607.00208","title":"SLIM-RL: Risk-Budgeted Random-Masking RL for Diffusion LLMs Without Trajectory Slicing","display_name":"SLIM-RL: Risk-Budgeted Random-Masking RL for Diffusion LLMs Without Trajectory Slicing","publication_year":2026,"publication_date":"2026-06-30","ids":{"openalex":"https://openalex.org/W7167032472","doi":"https://doi.org/10.48550/arxiv.2607.00208"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2607.00208","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00208","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.2607.00208","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139943223","display_name":"Ruikang Zhao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhao, Ruikang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139934732","display_name":"Zhenting Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Zhenting","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5076692827","display_name":"He Gao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Han","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139872077","display_name":"Ligong Han","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Han, Ligong","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/T10028","display_name":"Topic Modeling","score":0.3140999972820282,"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"}},"topics":[{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.3140999972820282,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.12530000507831573,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.05790000036358833,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/trajectory","display_name":"Trajectory","score":0.5299000144004822},{"id":"https://openalex.org/keywords/commit","display_name":"Commit","score":0.5281999707221985},{"id":"https://openalex.org/keywords/schedule","display_name":"Schedule","score":0.47609999775886536},{"id":"https://openalex.org/keywords/block","display_name":"Block (permutation group theory)","score":0.4593999981880188},{"id":"https://openalex.org/keywords/slicing","display_name":"Slicing","score":0.3970000147819519},{"id":"https://openalex.org/keywords/monotonic-function","display_name":"Monotonic function","score":0.3490999937057495},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.3253999948501587},{"id":"https://openalex.org/keywords/train","display_name":"Train","score":0.32249999046325684},{"id":"https://openalex.org/keywords/autoregressive-model","display_name":"Autoregressive model","score":0.3037000000476837}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6021000146865845},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.5299000144004822},{"id":"https://openalex.org/C153180980","wikidata":"https://www.wikidata.org/wiki/Q19776675","display_name":"Commit","level":2,"score":0.5281999707221985},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5242000222206116},{"id":"https://openalex.org/C68387754","wikidata":"https://www.wikidata.org/wiki/Q7271585","display_name":"Schedule","level":2,"score":0.47609999775886536},{"id":"https://openalex.org/C2777210771","wikidata":"https://www.wikidata.org/wiki/Q4927124","display_name":"Block (permutation group theory)","level":2,"score":0.4593999981880188},{"id":"https://openalex.org/C2776190703","wikidata":"https://www.wikidata.org/wiki/Q488148","display_name":"Slicing","level":2,"score":0.3970000147819519},{"id":"https://openalex.org/C72169020","wikidata":"https://www.wikidata.org/wiki/Q194404","display_name":"Monotonic function","level":2,"score":0.3490999937057495},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.3361000120639801},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.3253999948501587},{"id":"https://openalex.org/C190839683","wikidata":"https://www.wikidata.org/wiki/Q2448197","display_name":"Train","level":2,"score":0.32249999046325684},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.3037000000476837},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.30230000615119934},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.3019999861717224},{"id":"https://openalex.org/C52740198","wikidata":"https://www.wikidata.org/wiki/Q1539564","display_name":"Importance sampling","level":3,"score":0.298799991607666},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.29499998688697815},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.29499998688697815},{"id":"https://openalex.org/C2777402240","wikidata":"https://www.wikidata.org/wiki/Q6783436","display_name":"Masking (illustration)","level":2,"score":0.29109999537467957},{"id":"https://openalex.org/C2776654903","wikidata":"https://www.wikidata.org/wiki/Q2601463","display_name":"SAFER","level":2,"score":0.27959999442100525},{"id":"https://openalex.org/C41431624","wikidata":"https://www.wikidata.org/wiki/Q1053357","display_name":"Block size","level":3,"score":0.2786000072956085},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.27379998564720154},{"id":"https://openalex.org/C7252396","wikidata":"https://www.wikidata.org/wiki/Q1269210","display_name":"Dynode","level":4,"score":0.2662999927997589},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.2662999927997589},{"id":"https://openalex.org/C165464430","wikidata":"https://www.wikidata.org/wiki/Q1570441","display_name":"Parameterized complexity","level":2,"score":0.2662000060081482},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.26170000433921814},{"id":"https://openalex.org/C206729178","wikidata":"https://www.wikidata.org/wiki/Q2271896","display_name":"Scheduling (production processes)","level":2,"score":0.2612000107765198},{"id":"https://openalex.org/C147297375","wikidata":"https://www.wikidata.org/wiki/Q6674930","display_name":"Look-ahead","level":2,"score":0.2556000053882599},{"id":"https://openalex.org/C4679612","wikidata":"https://www.wikidata.org/wiki/Q866298","display_name":"Aggregate (composite)","level":2,"score":0.25540000200271606}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2607.00208","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00208","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.2607.00208","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2607.00208","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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reinforcement":[0],"learning":[1],"for":[2],"diffusion":[3],"large":[4],"language":[5],"models":[6],"(dLLMs)":[7],"has":[8],"largely":[9],"moved":[10],"to":[11,45],"trajectory-aware":[12],"methods.":[13],"The":[14,206,216],"current":[15],"state":[16],"of":[17,78],"the":[18,28,55,69,75,91,168,172,189],"art,":[19],"TraceRL,":[20],"holds":[21],"that":[22,35,52,61,107,128],"random":[23],"masking":[24,118],"is":[25,219],"mismatched":[26],"with":[27,54,82,102],"model's":[29],"inference":[30],"trajectory,":[31],"and":[32,156,175,202,214],"it":[33,194],"reconstructs":[34],"trajectory":[36],"during":[37],"training":[38,48,92,143],"by":[39,152,182,198],"slicing":[40],"each":[41,79],"rollout":[42,80],"into":[43],"up":[44],"K/s":[46],"trajectory-aligned":[47],"samples,":[49],"a":[50,83,103,121],"cost":[51],"grows":[53],"block":[56,146,165],"size":[57,147,166],"K.":[58],"We":[59],"show":[60],"this":[62],"mismatch":[63],"can":[64],"be":[65],"mitigated":[66],"without":[67],"reconstructing":[68],"trajectory.":[70],"Our":[71],"method,":[72],"SLIM-RL,":[73],"bounds":[74],"commit":[76,88],"risk":[77,89],"step":[81],"tau-budget":[84,207],"decoder,":[85],"reducing":[86],"aggregate":[87],"in":[90],"data.":[93],"During":[94],"optimization,":[95],"SLIM-RL":[96,133,170],"trains":[97],"on":[98,139,154,158,178,184,200,204],"these":[99],"risk-controlled":[100],"rollouts":[101],"trace-free":[104],"random-masking":[105],"objective":[106],"adapts":[108],"variance-reduction":[109],"tools,":[110],"combining":[111],"sequence-level":[112],"importance":[113],"sampling,":[114],"deterministic":[115],"quadrature":[116],"over":[117,150,196],"levels":[119],"under":[120,160],"mean-preserving,":[122],"monotonically":[123],"decreasing":[124],"per-block":[125],"mask":[126],"schedule":[127],"we":[129],"introduce.":[130],"On":[131,192],"SDAR-4B,":[132],"matches":[134],"TraceRL's":[135],"best":[136],"MATH500":[137,155,185],"accuracy":[138],"only":[140],"0.46x":[141],"its":[142],"samples":[144],"at":[145,221],"16,":[148],"improving":[149],"TraceRL":[151,197],"6.32%":[153],"11.05%":[157],"GSM8K":[159],"matched":[161],"dynamic":[162],"sampling.":[163],"At":[164],"4,":[167],"4B":[169],"surpasses":[171],"larger":[173],"LLaDA-8B":[174,181],"Dream-7B":[176],"dLLMs":[177],"math,":[179],"exceeding":[180],"10.76%":[183],"while":[186],"staying":[187],"below":[188],"autoregressive":[190],"Qwen2.5-7B.":[191],"code,":[193],"improves":[195],"4.20%":[199],"MBPP":[201],"3.65%":[203],"HumanEval.":[205],"decoder":[208],"transfers":[209],"training-free":[210],"across":[211],"LLaDA,":[212],"Dream,":[213],"SDAR.":[215],"source":[217],"code":[218],"available":[220],"https://github.com/laolaorkkkkk/SLIM-RL":[222],".":[223]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-07-03T00:00:00"}