{"id":"https://openalex.org/W7148301775","doi":"https://doi.org/10.48550/arxiv.2604.00886","title":"PixelPrune: Pixel-Level Adaptive Visual Token Reduction via Predictive Coding","display_name":"PixelPrune: Pixel-Level Adaptive Visual Token Reduction via Predictive Coding","publication_year":2026,"publication_date":"2026-04-01","ids":{"openalex":"https://openalex.org/W7148301775","doi":"https://doi.org/10.48550/arxiv.2604.00886"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.00886","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.00886","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":"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"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2604.00886","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5132823801","display_name":"Nan Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Nan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132762244","display_name":"Zhiwei Jin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Zhiwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132795805","display_name":"Chen Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Chen","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5132799166","display_name":"Haonan Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Haonan","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":0.4253999888896942,"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":0.4253999888896942,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.28209999203681946,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.09139999747276306,"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/encoder","display_name":"Encoder","score":0.5637999773025513},{"id":"https://openalex.org/keywords/speedup","display_name":"Speedup","score":0.5332000255584717},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5196999907493591},{"id":"https://openalex.org/keywords/exploit","display_name":"Exploit","score":0.49619999527931213},{"id":"https://openalex.org/keywords/lossy-compression","display_name":"Lossy compression","score":0.49380001425743103},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.48410001397132874},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.4343999922275543},{"id":"https://openalex.org/keywords/lossless-compression","display_name":"Lossless compression","score":0.42829999327659607},{"id":"https://openalex.org/keywords/coding","display_name":"Coding (social sciences)","score":0.37959998846054077}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.817799985408783},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.5637999773025513},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.5332000255584717},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5196999907493591},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4966999888420105},{"id":"https://openalex.org/C165696696","wikidata":"https://www.wikidata.org/wiki/Q11287","display_name":"Exploit","level":2,"score":0.49619999527931213},{"id":"https://openalex.org/C165021410","wikidata":"https://www.wikidata.org/wiki/Q55564","display_name":"Lossy compression","level":2,"score":0.49380001425743103},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.48410001397132874},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.4343999922275543},{"id":"https://openalex.org/C81081738","wikidata":"https://www.wikidata.org/wiki/Q55542","display_name":"Lossless compression","level":3,"score":0.42829999327659607},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.37959998846054077},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3684000074863434},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.35749998688697815},{"id":"https://openalex.org/C101738243","wikidata":"https://www.wikidata.org/wiki/Q786435","display_name":"Autoencoder","level":3,"score":0.34549999237060547},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.3425999879837036},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.33809998631477356},{"id":"https://openalex.org/C78548338","wikidata":"https://www.wikidata.org/wiki/Q2493","display_name":"Data compression","level":2,"score":0.328000009059906},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.32600000500679016},{"id":"https://openalex.org/C13481523","wikidata":"https://www.wikidata.org/wiki/Q412438","display_name":"Image compression","level":4,"score":0.3206999897956848},{"id":"https://openalex.org/C160633673","wikidata":"https://www.wikidata.org/wiki/Q355198","display_name":"Pixel","level":2,"score":0.3125},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.30880001187324524},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.299699991941452},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.29120001196861267},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.28870001435279846},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.28780001401901245},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.2847000062465668},{"id":"https://openalex.org/C57890076","wikidata":"https://www.wikidata.org/wiki/Q4680725","display_name":"Adaptive coding","level":4,"score":0.2775000035762787},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2773999869823456},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.27239999175071716},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.2578999996185303},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.25679999589920044},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.25609999895095825}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.00886","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.00886","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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2604.00886","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.00886","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":"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"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Document":[0],"understanding":[1],"and":[2,23,49,108,125,142,144,161],"GUI":[3,50,145],"interaction":[4],"are":[5,57],"among":[6],"the":[7,59,68,86,105,109,113],"highest-value":[8],"applications":[9],"of":[10,33,35,54,64],"Vision-Language":[11],"Models":[12],"(VLMs),":[13],"yet":[14],"they":[15],"impose":[16],"exceptionally":[17],"heavy":[18],"computational":[19],"burden:":[20],"fine-grained":[21],"text":[22],"small":[24],"UI":[25],"elements":[26],"demand":[27],"high-resolution":[28],"inputs":[29],"that":[30,40,148],"produce":[31],"tens":[32],"thousands":[34],"visual":[36],"tokens.":[37],"We":[38,71],"observe":[39],"this":[41,76],"cost":[42],"is":[43,119,166],"largely":[44],"wasteful":[45],"--":[46],"across":[47,138],"document":[48,143],"benchmarks,":[51],"only":[52],"22--71\\%":[53],"image":[55],"patches":[56,84],"pixel-unique,":[58],"rest":[60],"being":[61],"exact":[62],"duplicates":[63],"another":[65],"patch":[66],"in":[67,94],"same":[69],"image.":[70],"propose":[72],"\\textbf{PixelPrune},":[73],"which":[74],"exploits":[75],"pixel-level":[77],"redundancy":[78],"through":[79],"predictive-coding-based":[80],"compression,":[81],"pruning":[82],"redundant":[83],"\\emph{before}":[85],"Vision":[87],"Transformer":[88],"(ViT)":[89],"encoder.":[90],"Because":[91],"it":[92],"operates":[93],"pixel":[95],"space":[96],"prior":[97],"to":[98,157],"any":[99],"neural":[100],"computation,":[101],"PixelPrune":[102,149],"accelerates":[103],"both":[104],"ViT":[106],"encoder":[107],"downstream":[110],"LLM,":[111],"covering":[112],"full":[114],"inference":[115,159],"pipeline.":[116],"The":[117],"method":[118],"training-free,":[120],"requires":[121],"no":[122],"learnable":[123],"parameters,":[124],"supports":[126],"pixel-lossless":[127],"compression":[128,135],"($\u03c4{=}0$)":[129],"as":[130,132],"well":[131],"controlled":[133],"lossy":[134],"($\u03c4{>}0$).":[136],"Experiments":[137],"three":[139],"model":[140],"scales":[141],"benchmarks":[146],"show":[147],"maintains":[150],"competitive":[151],"task":[152],"accuracy":[153],"while":[154],"delivering":[155],"up":[156],"4.2$\\times$":[158],"speedup":[160],"1.9$\\times$":[162],"training":[163],"acceleration.":[164],"Code":[165],"available":[167],"at":[168],"https://github.com/OPPO-Mente-Lab/PixelPrune.":[169]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-03T00:00:00"}