{"id":"https://openalex.org/W7166662053","doi":"https://doi.org/10.48550/arxiv.2606.30345","title":"DRIFT: Difficulty Routing Self-DIstillation with Rhythm-Gated Exploration and Success BuFfer Training","display_name":"DRIFT: Difficulty Routing Self-DIstillation with Rhythm-Gated Exploration and Success BuFfer Training","publication_year":2026,"publication_date":"2026-06-29","ids":{"openalex":"https://openalex.org/W7166662053","doi":"https://doi.org/10.48550/arxiv.2606.30345"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.30345","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30345","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.2606.30345","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5048454873","display_name":"Haisen Luo","orcid":"https://orcid.org/0000-0002-2922-4919"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Luo, Haisen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139698299","display_name":"Yiwei Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Yiwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139635458","display_name":"Haoning Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Haoning","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139677182","display_name":"Dan Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Dan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5089766770","display_name":"Junxi Yin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Junxi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139708006","display_name":"Haotian Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Haotian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139693771","display_name":"Lei Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Lei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139656312","display_name":"Xiaoyu Tian","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tian, Xiaoyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139708752","display_name":"Shuaiting Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Shuaiting","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112525685","display_name":"Yuansheng Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Yuansheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5112411045","display_name":"Baoyan Guo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Guo, Baoyan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139676650","display_name":"Xiongfei Yan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yan, Xiongfei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122931632","display_name":"B Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Bolan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139674852","display_name":"Chengwei Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Chengwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139685911","display_name":"Ming Cui","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cui, Ming","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5101605878","display_name":"Jiong Chen","orcid":"https://orcid.org/0000-0002-9411-1689"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Jiong","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":0.19670000672340393,"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":0.19670000672340393,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.18209999799728394,"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":0.1356000006198883,"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":0.7728999853134155},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.5992000102996826},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.5978000164031982},{"id":"https://openalex.org/keywords/routing","display_name":"Routing (electronic design automation)","score":0.511900007724762},{"id":"https://openalex.org/keywords/concept-drift","display_name":"Concept drift","score":0.49709999561309814},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.4187999963760376},{"id":"https://openalex.org/keywords/tracking","display_name":"Tracking (education)","score":0.40959998965263367},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.39739999175071716}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7728999853134155},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7594000101089478},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.5992000102996826},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.5978000164031982},{"id":"https://openalex.org/C74172769","wikidata":"https://www.wikidata.org/wiki/Q1446839","display_name":"Routing (electronic design automation)","level":2,"score":0.511900007724762},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5088000297546387},{"id":"https://openalex.org/C60777511","wikidata":"https://www.wikidata.org/wiki/Q3045002","display_name":"Concept drift","level":3,"score":0.49709999561309814},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.47780001163482666},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.4187999963760376},{"id":"https://openalex.org/C2775936607","wikidata":"https://www.wikidata.org/wiki/Q466845","display_name":"Tracking (education)","level":2,"score":0.40959998965263367},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.39739999175071716},{"id":"https://openalex.org/C48103436","wikidata":"https://www.wikidata.org/wiki/Q599031","display_name":"State (computer science)","level":2,"score":0.3808000087738037},{"id":"https://openalex.org/C132758656","wikidata":"https://www.wikidata.org/wiki/Q5307365","display_name":"Dreyfus model of skill acquisition","level":2,"score":0.31220000982284546},{"id":"https://openalex.org/C77967617","wikidata":"https://www.wikidata.org/wiki/Q4677561","display_name":"Active learning (machine learning)","level":2,"score":0.3116999864578247},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.31040000915527344},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.30790001153945923},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.28529998660087585},{"id":"https://openalex.org/C106189395","wikidata":"https://www.wikidata.org/wiki/Q176789","display_name":"Markov decision process","level":3,"score":0.2651999890804291},{"id":"https://openalex.org/C24856439","wikidata":"https://www.wikidata.org/wiki/Q352483","display_name":"Adaptive routing","level":5,"score":0.2606000006198883},{"id":"https://openalex.org/C47177190","wikidata":"https://www.wikidata.org/wiki/Q207137","display_name":"Curriculum","level":2,"score":0.25450000166893005},{"id":"https://openalex.org/C74672266","wikidata":"https://www.wikidata.org/wiki/Q815859","display_name":"Language acquisition","level":2,"score":0.25189998745918274}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.30345","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30345","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.2606.30345","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30345","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":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Enabling":[0],"large":[1,72],"language":[2,73],"models":[3],"to":[4,53],"achieve":[5],"stable":[6,153],"self-improvement":[7,79],"without":[8],"external":[9],"expert":[10],"supervision":[11,47],"remains":[12],"a":[13,129,133,198,221],"central":[14],"challenge":[15],"in":[16],"complex":[17],"reasoning":[18,124],"tasks.":[19],"Existing":[20],"self-distillation":[21,105],"and":[22,34,51,88,102,106,132,160,172,193,216,224],"reinforcement":[23,107],"learning":[24,32,96,108,136],"methods":[25],"lack":[26],"explicit":[27],"mechanisms":[28],"for":[29,71],"tracking":[30],"problem-level":[31],"progress":[33],"adapting":[35],"optimization":[36,69],"strategies":[37],"accordingly.":[38],"Consequently,":[39],"training":[40],"may":[41],"over-optimize":[42],"easy":[43],"problems,":[44,50],"receive":[45],"weak":[46],"from":[48,148],"hard":[49],"fail":[52],"sufficiently":[54],"explore":[55],"borderline":[56],"cases.":[57],"To":[58],"resolve":[59],"these":[60],"issues,":[61],"we":[62],"propose":[63],"DRIFT,":[64],"an":[65,207],"online":[66],"self-evolution":[67],"policy":[68,114,154],"framework":[70],"models.":[74],"DRIFT":[75,138,164,186,205],"regulates":[76],"the":[77,82,94,99,111,117,146,166,179,183],"model's":[78,95],"process":[80],"through":[81],"joint":[83],"use":[84],"of":[85,169,209],"Difficulty":[86],"Routing":[87],"Rhythm":[89],"Gating.":[90],"The":[91],"former":[92],"identifies":[93],"state":[97],"at":[98,116],"problem":[100],"level":[101],"dynamically":[103],"allocates":[104],"signals,":[109],"while":[110,143],"latter":[112],"refines":[113],"updates":[115],"token":[118],"level,":[119],"concentrating":[120],"exploration":[121],"on":[122,203],"critical":[123],"positions.":[125],"By":[126],"further":[127],"incorporating":[128],"success":[130],"buffer":[131],"two-stage":[134],"curriculum":[135],"strategy,":[137],"preserves":[139],"high-quality":[140],"historical":[141],"experience":[142],"progressively":[144],"guiding":[145],"model":[147,162],"reliable":[149],"behavior":[150],"acquisition":[151],"toward":[152],"evolution.":[155],"Evaluated":[156],"across":[157,174],"five":[158,184],"benchmarks":[159],"three":[161],"scales,":[163],"surpasses":[165],"peak":[167],"performance":[168],"both":[170],"GRPO":[171,190,213],"SDPO":[173,194,217],"all":[175,227],"evaluated":[176],"metrics.":[177],"On":[178],"average":[180],"score":[181],"over":[182,212],"benchmarks,":[185],"achieves":[187],"79.5$\\%$,":[188],"outperforming":[189,226],"by":[191,195,214,218],"9.5$\\%$":[192],"7.5$\\%$,":[196],"establishing":[197],"new":[199,222],"state-of-the-art":[200,223],"result.":[201],"Notably,":[202],"ToolUse,":[204],"reaches":[206],"accuracy":[208],"79.2$\\%$,":[210],"improving":[211],"13.5$\\%$":[215],"10.7$\\%$,":[219],"setting":[220],"substantially":[225],"concurrent":[228],"methods.":[229]},"counts_by_year":[],"updated_date":"2026-07-01T06:29:00.853634","created_date":"2026-07-01T00:00:00"}