{"id":"https://openalex.org/W7160913388","doi":"https://doi.org/10.48550/arxiv.2605.10050","title":"EchoPrune: Interpreting Redundancy as Temporal Echoes for Efficient VideoLLMs","display_name":"EchoPrune: Interpreting Redundancy as Temporal Echoes for Efficient VideoLLMs","publication_year":2026,"publication_date":"2026-05-11","ids":{"openalex":"https://openalex.org/W7160913388","doi":"https://doi.org/10.48550/arxiv.2605.10050"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.10050","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10050","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.2605.10050","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5053783274","display_name":"Jiameng Li","orcid":"https://orcid.org/0000-0002-3486-3758"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Jiameng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5085836060","display_name":"Minye Wu","orcid":"https://orcid.org/0000-0002-8163-9513"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Minye","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5090145580","display_name":"Jiezhang Cao","orcid":"https://orcid.org/0000-0002-0050-4714"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Cao, Jiezhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135970097","display_name":"Aleksei Tiulpin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tiulpin, Aleksei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135936186","display_name":"Matthew B. Blaschko","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Blaschko, Matthew B.","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":0.9368000030517578,"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":0.9368000030517578,"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":0.025800000876188278,"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":0.01080000028014183,"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/security-token","display_name":"Security token","score":0.6571000218391418},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4406000077724457},{"id":"https://openalex.org/keywords/redundancy","display_name":"Redundancy (engineering)","score":0.43290001153945923},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.40709999203681946},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.3734999895095825},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.36410000920295715},{"id":"https://openalex.org/keywords/process","display_name":"Process (computing)","score":0.35370001196861267},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.3425999879837036}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8001999855041504},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.6571000218391418},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.60589998960495},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.4763999879360199},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4406000077724457},{"id":"https://openalex.org/C152124472","wikidata":"https://www.wikidata.org/wiki/Q1204361","display_name":"Redundancy (engineering)","level":2,"score":0.43290001153945923},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.40709999203681946},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.3734999895095825},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.36410000920295715},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.3619000017642975},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.35370001196861267},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.3425999879837036},{"id":"https://openalex.org/C68339613","wikidata":"https://www.wikidata.org/wiki/Q1549489","display_name":"Speedup","level":2,"score":0.33379998803138733},{"id":"https://openalex.org/C2778738651","wikidata":"https://www.wikidata.org/wiki/Q16546687","display_name":"Novelty","level":2,"score":0.32100000977516174},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.3160000145435333},{"id":"https://openalex.org/C119666444","wikidata":"https://www.wikidata.org/wiki/Q5977280","display_name":"Temporal resolution","level":2,"score":0.31459999084472656},{"id":"https://openalex.org/C126042441","wikidata":"https://www.wikidata.org/wiki/Q1324888","display_name":"Frame (networking)","level":2,"score":0.30070000886917114},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.2948000133037567},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.2799000144004822},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.27239999175071716},{"id":"https://openalex.org/C77277458","wikidata":"https://www.wikidata.org/wiki/Q1969246","display_name":"Temporal database","level":2,"score":0.27149999141693115},{"id":"https://openalex.org/C158154518","wikidata":"https://www.wikidata.org/wiki/Q7310970","display_name":"Relevance (law)","level":2,"score":0.2678000032901764},{"id":"https://openalex.org/C113954288","wikidata":"https://www.wikidata.org/wiki/Q186885","display_name":"Timestamp","level":2,"score":0.26440000534057617},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.2513999938964844}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.10050","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10050","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.2605.10050","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10050","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":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Long-form":[0],"video":[1,91,185],"understanding":[2,186],"remains":[3],"challenging":[4],"for":[5,212],"Video":[6],"Large":[7],"Language":[8],"Models":[9],"(VideoLLMs),":[10],"as":[11,42,93],"the":[12,103,173,200],"dense":[13],"frame":[14],"sampling":[15,22],"introduces":[16],"massive":[17],"visual":[18,81,122,130],"tokens":[19,92,131,154],"while":[20,161],"sparse":[21],"risks":[23],"missing":[24],"critical":[25],"temporal":[26,75,94,139,159],"evidence":[27],"and":[28,56,68,137,146,158,181,208],"leading":[29],"to":[30,88,167,193,196],"LLM":[31],"hallucination.":[32],"Existing":[33],"training-free":[34,69],"token":[35,70,82,98,202],"reduction":[36],"methods":[37],"either":[38],"treat":[39],"videos":[40],"equally":[41],"static":[43],"images":[44],"or":[45,120],"rely":[46],"on":[47,125,178,214],"segment-level":[48],"merging":[49],"heuristics,":[50],"which":[51],"weaken":[52],"fine-grained":[53],"spatiotemporal":[54],"modeling":[55],"introduce":[57],"additional":[58],"overhead.":[59],"In":[60],"this":[61,126],"paper,":[62],"we":[63],"propose":[64],"EchoPrune,":[65],"a":[66,78,97,109],"lightweight":[67],"pruning":[71],"method":[72],"that":[73,189],"improves":[74],"resolution":[76],"under":[77,199],"fixed":[79],"LLM-side":[80],"budget.":[83,175],"Our":[84],"core":[85],"idea":[86],"is":[87,99,107],"interpret":[89],"redundant":[90,111],"echoes:":[95],"if":[96],"well":[100],"reconstructed":[101],"from":[102],"previous":[104],"frame,":[105],"it":[106,114],"merely":[108],"temporally":[110],"echo;":[112],"otherwise,":[113],"may":[115],"capture":[116],"new":[117],"events,":[118],"motion,":[119],"query-relevant":[121],"evidence.":[123],"Based":[124],"insight,":[127],"EchoPrune":[128,190],"scores":[129],"by":[132,143],"(i)":[133],"query-guided":[134],"crossmodal":[135],"relevance":[136],"(ii)":[138],"reconstruction":[140],"error,":[141],"measured":[142],"correspondence":[144],"matching":[145,148],"echo":[147],"across":[149,183],"consecutive":[150],"frames.":[151],"The":[152],"selected":[153],"preserve":[155],"task-relevant":[156],"cues":[157],"novelty":[160],"suppressing":[162],"predictable":[163],"redundancy,":[164],"allowing":[165],"VideoLLMs":[166,192],"observe":[168],"more":[169],"frames":[170,198],"without":[171],"increasing":[172],"decoding":[174],"Extensive":[176],"experiments":[177],"LLaVA-OV,":[179],"Qwen2.5VL,":[180],"Qwen3VL":[182],"six":[184],"benchmarks":[187],"show":[188],"enables":[191],"process":[194],"up":[195],"20x":[197],"same":[201],"budget,":[203],"yielding":[204],"improved":[205],"performance":[206],"(+8.6%)":[207],"inference":[209],"speedup":[210],"(5.6x":[211],"prefilling)":[213],"Qwen2.5VL-7B.":[215]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-13T00:00:00"}