{"id":"https://openalex.org/W4404780907","doi":"https://doi.org/10.18653/v1/2024.findings-emnlp.754","title":"Enhancing Alignment using Curriculum Learning & Ranked Preferences","display_name":"Enhancing Alignment using Curriculum Learning & Ranked Preferences","publication_year":2024,"publication_date":"2024-01-01","ids":{"openalex":"https://openalex.org/W4404780907","doi":"https://doi.org/10.18653/v1/2024.findings-emnlp.754"},"language":"en","primary_location":{"id":"doi:10.18653/v1/2024.findings-emnlp.754","doi":"https://doi.org/10.18653/v1/2024.findings-emnlp.754","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.findings-emnlp.754","pdf_url":"https://aclanthology.org/2024.findings-emnlp.754.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2024","raw_type":"proceedings-article"},"volume":null,"type":"conference-paper","is_traditional":true,"indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://aclanthology.org/2024.findings-emnlp.754.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5104257880","display_name":"Pulkit Pattnaik","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pulkit Pattnaik","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086568118","display_name":"Rishabh Maheshwary","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Rishabh Maheshwary","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088147173","display_name":"Kelechi Ogueji","orcid":"https://orcid.org/0009-0004-7919-9303","observed_orcids":["https://orcid.org/0009-0004-7919-9303"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kelechi Ogueji","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101500443","display_name":"Vikas Yadav","orcid":"https://orcid.org/0000-0002-8536-2822","observed_orcids":["https://orcid.org/0000-0002-8536-2822"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vikas Yadav","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5016214507","display_name":"Sathwik Tejaswi Madhusudhan","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sathwik Tejaswi Madhusudhan","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":1.6796,"has_fulltext":true,"cited_by_count":11,"citation_normalized_percentile":{"value":0.87452872,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":97,"max":99},"biblio":{"volume":null,"issue":null,"first_page":"12891","last_page":"12907"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.2754000127315521,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.2754000127315521,"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":0.25,"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":0.1152999997138977,"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/direct-preference-optimization","display_name":"direct preference optimization","score":0.6629999876022339},{"id":"https://openalex.org/keywords/curriculum-learning","display_name":"curriculum learning","score":0.8989999890327454},{"id":"https://openalex.org/keywords/preference-pairs","display_name":"preference pairs","score":0.6970000267028809},{"id":"https://openalex.org/keywords/large-language-model-alignment","display_name":"large language model alignment","score":0.8980000019073486},{"id":"https://openalex.org/keywords/human-preference-data","display_name":"human preference data","score":0.7990000247955322},{"id":"https://openalex.org/keywords/mt-bench","display_name":"MT-Bench","score":0.5360000133514404},{"id":"https://openalex.org/keywords/vicuna","display_name":"vicu\u00f1a","score":0.7049999833106995}],"concepts":[{"id":"https://openalex.org/C47177190","wikidata":"https://www.wikidata.org/wiki/Q207137","display_name":"Curriculum","level":2,"score":0.6547539830207825},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6350573301315308},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.37538933753967285},{"id":"https://openalex.org/C145420912","wikidata":"https://www.wikidata.org/wiki/Q853077","display_name":"Mathematics education","level":1,"score":0.36446714401245117},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3573489487171173},{"id":"https://openalex.org/C56739046","wikidata":"https://www.wikidata.org/wiki/Q192060","display_name":"Knowledge management","level":1,"score":0.35343846678733826},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.18993088603019714},{"id":"https://openalex.org/C19417346","wikidata":"https://www.wikidata.org/wiki/Q7922","display_name":"Pedagogy","level":1,"score":0.1320626139640808}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2024.findings-emnlp.754","doi":"https://doi.org/10.18653/v1/2024.findings-emnlp.754","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.findings-emnlp.754","pdf_url":"https://aclanthology.org/2024.findings-emnlp.754.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2024","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2024.findings-emnlp.754","doi":"https://doi.org/10.18653/v1/2024.findings-emnlp.754","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2024.findings-emnlp.754","pdf_url":"https://aclanthology.org/2024.findings-emnlp.754.pdf","source":{"id":"https://openalex.org/S7407096085","display_name":"Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/P4310320244","host_organization_name":"Association for Computational Linguistics","host_organization_lineage":["https://openalex.org/P4310320244"],"host_organization_lineage_names":["Association for Computational Linguistics"],"type":"conference"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Findings of the Association for Computational Linguistics: EMNLP 2024","raw_type":"proceedings-article"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4404780907.pdf","grobid_xml":"https://content.openalex.org/works/W4404780907.grobid-xml"},"referenced_works_count":25,"referenced_works":["https://openalex.org/W2736601468","https://openalex.org/W2798047776","https://openalex.org/W4223908421","https://openalex.org/W4365211596","https://openalex.org/W4366341216","https://openalex.org/W4367000491","https://openalex.org/W4377121527","https://openalex.org/W4386437475","https://openalex.org/W4387322976","https://openalex.org/W4387561528","https://openalex.org/W4387595881","https://openalex.org/W4387801352","https://openalex.org/W4387994989","https://openalex.org/W4388890574","https://openalex.org/W4389519291","https://openalex.org/W4390437503","https://openalex.org/W4390573302","https://openalex.org/W4390723197","https://openalex.org/W4391047180","https://openalex.org/W4391555989","https://openalex.org/W4391591751","https://openalex.org/W4391799209","https://openalex.org/W4393335839","https://openalex.org/W4402670165","https://openalex.org/W7133208539"],"related_works":["https://openalex.org/W4392781846","https://openalex.org/W4417251511","https://openalex.org/W7166832820","https://openalex.org/W7127351271","https://openalex.org/W4407806932","https://openalex.org/W7196938767","https://openalex.org/W4404400430","https://openalex.org/W4405173242","https://openalex.org/W4391988381","https://openalex.org/W4415795724"],"abstract_inverted_index":{"Direct":[0],"Preference":[1],"Optimization":[2],"(DPO)":[3],"is":[4,100],"aneffective":[5],"technique":[6],"that":[7,122],"leverages":[8],"pairwise":[9],"preference":[10,49,64,84,125,182],"data":[11],"(one":[12],"chosen":[13],"and":[14,66,156,162],"rejected":[15],"response":[16],"per":[17],"prompt)":[18],"to":[19,22,38,42,46,76,88,91,102,173,177],"align":[20],"LLMs":[21,59,142],"human":[23],"preferences.In":[24],"practice,":[25],"multiple":[26,48,63,83],"responses":[27,45],"could":[28],"exist":[29],"for":[30,51],"a":[31,52,70,130],"given":[32,53],"prompt":[33],"with":[34,136,143,168],"varying":[35],"quality":[36],"relative":[37],"each":[39],"other.We":[40],"propose":[41],"utilize":[43],"these":[44],"create":[47],"pairs":[50,65,85,183],"prompt.Our":[54],"work":[55],"focuses":[56],"on":[57,110,134,153],"aligning":[58],"by":[60],"systematically":[61],"curating":[62],"presenting":[67],"them":[68],"in":[69,165,185],"meaningful":[71],"manner":[72],"facilitating":[73],"curriculum":[74,96],"learning":[75],"enhance":[77],"the":[78,149,181],"prominent":[79],"DPO":[80,120],"technique.We":[81],"order":[82],"from":[86],"easy":[87],"hard,":[89],"according":[90],"various":[92],"criteria":[93],"thus":[94],"emulating":[95],"learning.Our":[97],"method,":[98],"which":[99],"referred":[101],"as":[103],"Curri-DPO":[104,128],"consistently":[105],"shows":[106],"increased":[107],"performance":[108],"gains":[109,170],"MTbench,":[111],"Vicuna":[112],"bench,":[113],"WizardLM,":[114,155],"highlighting":[115],"its":[116],"effectiveness":[117],"over":[118],"standard":[119,178],"setting":[121],"utilizes":[123],"single":[124],"pair.More":[126],"specifically,":[127],"achieves":[129,148],"score":[131],"of":[132,140,171],"7.43":[133],"MT-bench":[135],"Zephyr-7B,":[137],"outperforming":[138],"majority":[139],"existing":[141],"similar":[144],"parameter":[145],"size.Curri-DPO":[146],"also":[147],"highest":[150],"win":[151],"rates":[152],"Vicuna,":[154],"UltraFeedback":[157],"test":[158],"sets":[159],"(90.7%,":[160],"87.1%,":[161],"87.9%":[163],"respectively)":[164],"our":[166],"experiments,":[167],"notable":[169],"up":[172],"7.5%":[174],"when":[175],"compared":[176],"DPO.We":[179],"release":[180],"used":[184],"alignment":[186],"at:":[187],"ServiceNow-":[188],"AI/Curriculum_DPO_preferences.":[189]},"counts_by_year":[{"year":2026,"cited_by_count":7},{"year":2025,"cited_by_count":4}],"updated_date":"2026-11-10T11:39:43.021295","created_date":"2024-11-28T00:00:00"}