{"id":"https://openalex.org/W4400375208","doi":"https://doi.org/10.48550/arxiv.2407.02466","title":"PWM: Policy Learning with Multi-Task World Models","display_name":"PWM: Policy Learning with Multi-Task World Models","publication_year":2024,"publication_date":"2024-07-02","ids":{"openalex":"https://openalex.org/W4400375208","doi":"https://doi.org/10.48550/arxiv.2407.02466"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2407.02466","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2407.02466","pdf_url":"https://arxiv.org/pdf/2407.02466","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2407.02466","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5073871707","display_name":"Ignat Georgiev","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I130701444","display_name":"Georgia Institute of Technology","ror":"https://ror.org/01zkghx44","country_code":"US","type":"education","lineage":["https://openalex.org/I130701444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Georgiev, Ignat","raw_affiliation_strings":["Georgia Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Georgia Institute of Technology","institution_ids":["https://openalex.org/I130701444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100029933","display_name":"Varun Giridhar","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I130701444","display_name":"Georgia Institute of Technology","ror":"https://ror.org/01zkghx44","country_code":"US","type":"education","lineage":["https://openalex.org/I130701444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Giridhar, Varun","raw_affiliation_strings":["Georgia Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Georgia Institute of Technology","institution_ids":["https://openalex.org/I130701444"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5086722753","display_name":"Nicklas Hansen","orcid":"https://orcid.org/0000-0001-9897-4003","observed_orcids":["https://orcid.org/0000-0001-9897-4003"]},"institutions":[{"id":"https://openalex.org/I36258959","display_name":"University of California San Diego","ror":"https://ror.org/0168r3w48","country_code":"US","type":"education","lineage":["https://openalex.org/I36258959"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Hansen, Nicklas","raw_affiliation_strings":["UC San Diego"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"UC San Diego","institution_ids":["https://openalex.org/I36258959"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5061193324","display_name":"Animesh Garg","orcid":"https://orcid.org/0000-0003-0482-4296","observed_orcids":["https://orcid.org/0000-0003-0482-4296"]},"institutions":[{"id":"https://openalex.org/I130701444","display_name":"Georgia Institute of Technology","ror":"https://ror.org/01zkghx44","country_code":"US","type":"education","lineage":["https://openalex.org/I130701444"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Garg, Animesh","raw_affiliation_strings":["Georgia Institute of Technology"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Georgia Institute of Technology","institution_ids":["https://openalex.org/I130701444"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":true,"cited_by_count":16,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":1.0,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":1.0,"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":0.0,"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/T10906","display_name":"AI-based Problem Solving and Planning","score":0.0,"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/policy-learning","display_name":"policy learning","score":0.574999988079071},{"id":"https://openalex.org/keywords/world-models","display_name":"world models","score":0.9200000166893005},{"id":"https://openalex.org/keywords/multi-task-learning","display_name":"multi-task learning","score":0.8230000138282776},{"id":"https://openalex.org/keywords/model-based-reinforcement-learning","display_name":"model-based reinforcement learning","score":0.9409999847412109},{"id":"https://openalex.org/keywords/continuous-control","display_name":"continuous control","score":0.7549999952316284},{"id":"https://openalex.org/keywords/first-order-methods","display_name":"first-order methods","score":0.7089999914169312},{"id":"https://openalex.org/keywords/simulation-based-training","display_name":"simulation-based training","score":0.39800000190734863},{"id":"https://openalex.org/keywords/high-dimensional-action-spaces","display_name":"high-dimensional action spaces","score":0.6940000057220459}],"concepts":[{"id":"https://openalex.org/C2779436431","wikidata":"https://www.wikidata.org/wiki/Q30672407","display_name":"Policy learning","level":2,"score":0.534285306930542},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.50058913230896},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.22993403673171997}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2407.02466","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2407.02466","pdf_url":"https://arxiv.org/pdf/2407.02466","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2407.02466","doi":"https://doi.org/10.48550/arxiv.2407.02466","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2407.02466","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2407.02466","doi":null,"is_oa":true,"landing_page_url":"http://arxiv.org/abs/2407.02466","pdf_url":"https://arxiv.org/pdf/2407.02466","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,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4400375208.pdf","grobid_xml":"https://content.openalex.org/works/W4400375208.grobid-xml"},"referenced_works_count":15,"referenced_works":["https://openalex.org/W1757796397","https://openalex.org/W2736601468","https://openalex.org/W2904246096","https://openalex.org/W2992977009","https://openalex.org/W3037932933","https://openalex.org/W3107736070","https://openalex.org/W3127062143","https://openalex.org/W3195577433","https://openalex.org/W3213654397","https://openalex.org/W4283731125","https://openalex.org/W4307868887","https://openalex.org/W4381586832","https://openalex.org/W4387764390","https://openalex.org/W4402890475","https://openalex.org/W4409147637"],"related_works":["https://openalex.org/W4416766039","https://openalex.org/W4390438280","https://openalex.org/W4387171239","https://openalex.org/W7135200086","https://openalex.org/W7160501366","https://openalex.org/W4387964172","https://openalex.org/W7162522727","https://openalex.org/W7147336522","https://openalex.org/W7127331852","https://openalex.org/W4416585915"],"abstract_inverted_index":{"Reinforcement":[0],"Learning":[1],"(RL)":[2],"has":[3],"made":[4],"significant":[5],"strides":[6],"in":[7,12,110],"complex":[8],"tasks":[9,120],"but":[10,30,48],"struggles":[11],"multi-task":[13,79],"settings":[14],"with":[15,78,121],"different":[16],"embodiments.":[17],"World":[18,80],"model":[19,94],"methods":[20,37,44,129],"offer":[21],"scalability":[22],"by":[23],"learning":[24,77],"a":[25,83],"simulation":[26],"of":[27],"the":[28,66,92],"environment":[29],"often":[31],"rely":[32],"on":[33,97,152],"inefficient":[34],"gradient-free":[35],"optimization":[36,63,109],"for":[38,88],"policy":[39],"extraction.":[40],"In":[41],"contrast,":[42],"gradient-based":[43],"exhibit":[45],"lower":[46],"variance":[47],"fail":[49],"to":[50,123,137,143],"handle":[51],"discontinuities.":[52],"Our":[53],"work":[54],"reveals":[55],"that":[56,130],"well-regularized":[57],"world":[58,93],"models":[59],"can":[60],"generate":[61],"smoother":[62],"landscapes":[64],"than":[65,112,147],"actual":[67],"dynamics,":[68],"facilitating":[69],"more":[70],"effective":[71],"first-order":[72,108],"optimization.":[73],"We":[74],"introduce":[75],"Policy":[76],"Models":[81],"(PWM),":[82],"novel":[84],"model-based":[85],"RL":[86],"algorithm":[87],"continuous":[89],"control.":[90],"Initially,":[91],"is":[95],"pre-trained":[96],"offline":[98],"data,":[99],"and":[100,127,157],"then":[101],"policies":[102],"are":[103,159],"extracted":[104],"from":[105],"it":[106],"using":[107],"less":[111],"10":[113],"minutes":[114],"per":[115],"task.":[116],"PWM":[117,135],"effectively":[118],"solves":[119],"up":[122,142],"152":[124],"action":[125],"dimensions":[126],"outperforms":[128],"use":[131],"ground-truth":[132],"dynamics.":[133],"Additionally,":[134],"scales":[136],"an":[138],"80-task":[139],"setting,":[140],"achieving":[141],"27%":[144],"higher":[145],"rewards":[146],"existing":[148],"baselines":[149],"without":[150],"relying":[151],"costly":[153],"online":[154],"planning.":[155],"Visualizations":[156],"code":[158],"available":[160],"at":[161],"https://www.imgeorgiev.com/pwm/.":[162]},"counts_by_year":[{"year":2026,"cited_by_count":8},{"year":2025,"cited_by_count":7},{"year":2024,"cited_by_count":1}],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2024-07-06T00:00:00"}