{"id":"https://openalex.org/W7164020090","doi":"https://doi.org/10.48550/arxiv.2606.08641","title":"Learnable Token Sparsification for Efficient Gigapixel Whole Slide Image Reasoning","display_name":"Learnable Token Sparsification for Efficient Gigapixel Whole Slide Image Reasoning","publication_year":2026,"publication_date":"2026-06-07","ids":{"openalex":"https://openalex.org/W7164020090","doi":"https://doi.org/10.48550/arxiv.2606.08641"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.08641","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08641","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.2606.08641","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5129017581","display_name":"Jingzhi Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Jingzhi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128646065","display_name":"Landi He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Landi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138272235","display_name":"Zhuo Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Zhuo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5127926984","display_name":"Shawn Young","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Young, Shawn","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138210736","display_name":"Lijian Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Lijian","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":0.46560001373291016,"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":0.46560001373291016,"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":0.2468000054359436,"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":0.06870000064373016,"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/context","display_name":"Context (archaeology)","score":0.4952000081539154},{"id":"https://openalex.org/keywords/component","display_name":"Component (thermodynamics)","score":0.48750001192092896},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.4661000072956085},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.4564000070095062},{"id":"https://openalex.org/keywords/pruning","display_name":"Pruning","score":0.439300000667572},{"id":"https://openalex.org/keywords/upsampling","display_name":"Upsampling","score":0.4171999990940094},{"id":"https://openalex.org/keywords/monocular","display_name":"Monocular","score":0.41359999775886536},{"id":"https://openalex.org/keywords/occam","display_name":"occam","score":0.3977999985218048},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.39660000801086426},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.38190001249313354}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7875000238418579},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5187000036239624},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.4952000081539154},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.48750001192092896},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.4661000072956085},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.4564000070095062},{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.439300000667572},{"id":"https://openalex.org/C110384440","wikidata":"https://www.wikidata.org/wiki/Q1143270","display_name":"Upsampling","level":3,"score":0.4171999990940094},{"id":"https://openalex.org/C65909025","wikidata":"https://www.wikidata.org/wiki/Q1945033","display_name":"Monocular","level":2,"score":0.41359999775886536},{"id":"https://openalex.org/C78469957","wikidata":"https://www.wikidata.org/wiki/Q838062","display_name":"occam","level":2,"score":0.3977999985218048},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.39660000801086426},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.38190001249313354},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3813999891281128},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.3698999881744385},{"id":"https://openalex.org/C28855332","wikidata":"https://www.wikidata.org/wiki/Q198099","display_name":"Quantization (signal processing)","level":2,"score":0.3684999942779541},{"id":"https://openalex.org/C173801870","wikidata":"https://www.wikidata.org/wiki/Q201413","display_name":"Heuristic","level":2,"score":0.3684999942779541},{"id":"https://openalex.org/C157202957","wikidata":"https://www.wikidata.org/wiki/Q1659609","display_name":"Image warping","level":2,"score":0.3596000075340271},{"id":"https://openalex.org/C2778112365","wikidata":"https://www.wikidata.org/wiki/Q3511065","display_name":"Sequence (biology)","level":2,"score":0.35920000076293945},{"id":"https://openalex.org/C111335779","wikidata":"https://www.wikidata.org/wiki/Q3454686","display_name":"Reduction (mathematics)","level":2,"score":0.35670000314712524},{"id":"https://openalex.org/C101468663","wikidata":"https://www.wikidata.org/wiki/Q1620158","display_name":"Modular design","level":2,"score":0.33880001306533813},{"id":"https://openalex.org/C99498987","wikidata":"https://www.wikidata.org/wiki/Q2210247","display_name":"Noise (video)","level":3,"score":0.3328000009059906},{"id":"https://openalex.org/C137800194","wikidata":"https://www.wikidata.org/wiki/Q11713455","display_name":"Interpolation (computer graphics)","level":3,"score":0.32499998807907104},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3107999861240387},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.31060001254081726},{"id":"https://openalex.org/C28719098","wikidata":"https://www.wikidata.org/wiki/Q44946","display_name":"Point (geometry)","level":2,"score":0.3077000081539154},{"id":"https://openalex.org/C155846161","wikidata":"https://www.wikidata.org/wiki/Q1143367","display_name":"Graphical model","level":2,"score":0.30239999294281006},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.29670000076293945},{"id":"https://openalex.org/C118505674","wikidata":"https://www.wikidata.org/wiki/Q42586063","display_name":"Encoder","level":2,"score":0.29649999737739563},{"id":"https://openalex.org/C81081738","wikidata":"https://www.wikidata.org/wiki/Q55542","display_name":"Lossless compression","level":3,"score":0.29010000824928284},{"id":"https://openalex.org/C153258448","wikidata":"https://www.wikidata.org/wiki/Q1199743","display_name":"Gradient descent","level":3,"score":0.28999999165534973},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.2881999909877777},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.2865999937057495},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.2833999991416931},{"id":"https://openalex.org/C9417928","wikidata":"https://www.wikidata.org/wiki/Q1070689","display_name":"Image processing","level":3,"score":0.2800000011920929},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.2754000127315521},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.271699994802475},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.27129998803138733},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.26840001344680786},{"id":"https://openalex.org/C17020691","wikidata":"https://www.wikidata.org/wiki/Q139677","display_name":"Operator (biology)","level":5,"score":0.2646999955177307},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.26409998536109924}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.08641","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08641","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.2606.08641","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.08641","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":{"The":[0],"processing":[1],"of":[2,20,84,123,187,197,242],"gigapixel":[3,257],"whole":[4,66,258],"slide":[5,67,259],"images":[6,68],"within":[7],"vision":[8,216],"language":[9,217],"models":[10],"faces":[11],"a":[12,70,90,107,114,127,134,160,184,249],"major":[13],"difficulty":[14],"due":[15],"to":[16,77,165,183,255],"an":[17,79],"excessive":[18],"number":[19],"visual":[21,180,235],"tokens.":[22],"Existing":[23],"solutions":[24],"typically":[25],"rely":[26],"on":[27,207,226],"spatial":[28,144],"downsampling":[29],"or":[30],"heuristic":[31],"pruning":[32,101],"strategies":[33],"that":[34,118,138],"operate":[35],"without":[36,142],"training,":[37,104],"and":[38,155,214,229,238],"these":[39],"methods":[40],"often":[41],"discard":[42],"subtle":[43],"but":[44],"clinically":[45],"meaningful":[46],"patterns":[47],"because":[48],"pathological":[49],"evidence":[50],"is":[51,152],"scattered":[52],"irregularly":[53],"across":[54],"the":[55,75,99,120,149,156,168,179,198,234,240],"tissue.":[56],"To":[57,94],"overcome":[58],"this":[59,246],"limitation,":[60],"we":[61,105],"reformulate":[62],"token":[63],"reduction":[64],"in":[65],"as":[69,193,195],"trainable":[71],"sparsification":[72],"problem,":[73],"allowing":[74],"model":[76],"learn":[78],"optimal":[80],"selection":[81],"strategy":[82],"instead":[83],"following":[85],"fixed":[86],"heuristics.":[87],"We":[88],"propose":[89],"decoupled":[91],"routing":[92],"architecture.":[93],"enable":[95],"gradient":[96],"propagation":[97],"through":[98],"nondifferentiable":[100],"operation":[102],"during":[103],"introduce":[106],"component":[108,112],"called":[109],"SparseLearn.":[110],"This":[111],"uses":[113],"variance-preserving":[115],"noise":[116],"gate":[117],"regulates":[119],"information":[121],"flow":[122],"each":[124],"patch":[125],"via":[126],"differentiable":[128],"Soft":[129],"Top-K":[130,163],"operator,":[131],"together":[132],"with":[133],"diagonal":[135],"attention":[136],"denoiser":[137],"recovers":[139],"perturbed":[140],"representations":[141],"leaking":[143],"information.":[145],"At":[146],"inference":[147],"time,":[148],"SparseLearn":[150],"module":[151],"entirely":[153],"discarded,":[154],"trained":[157],"scorer":[158],"applies":[159],"deterministic":[161],"Hard":[162],"operator":[164],"keep":[166],"only":[167],"highest":[169],"scoring":[170],"32":[171,189],"tokens,":[172,190],"incurring":[173],"no":[174],"extra":[175],"computation.":[176],"By":[177,232],"compressing":[178],"sequence":[181],"down":[182],"sparse":[185,243],"set":[186],"just":[188],"which":[191],"represents":[192],"little":[194],"0.78%":[196],"original":[199],"length,":[200],"our":[201],"framework":[202],"achieves":[203],"73.32%":[204],"overall":[205],"accuracy":[206],"SlideBench":[208,227],"(TCGA),":[209],"consistently":[210],"surpassing":[211],"sampling-based":[212],"baselines":[213],"general-purpose":[215],"models.":[218],"It":[219],"also":[220],"demonstrates":[221],"strong":[222],"zero":[223],"shot":[224],"generalization":[225],"(BCNB)":[228],"WSI":[230],"VQA*.":[231],"resolving":[233],"context":[236],"bottleneck":[237],"preventing":[239],"dilution":[241],"diagnostic":[244],"evidence,":[245],"work":[247],"provides":[248],"highly":[250],"efficient":[251],"paradigm":[252],"for":[253],"end":[254,256],"image":[260],"reasoning.":[261]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-10T00:00:00"}