{"id":"https://openalex.org/W7124417543","doi":"https://doi.org/10.14711/thesis-hdl167718","title":"Enabling Explainable AI with Transformer Models: Opportunities and Limitations in Visual and Textual Concept Generation","display_name":"Enabling Explainable AI with Transformer Models: Opportunities and Limitations in Visual and Textual Concept Generation","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W7124417543","doi":"https://doi.org/10.14711/thesis-hdl167718"},"language":"en","primary_location":{"id":"doi:10.14711/thesis-hdl167718","doi":"https://doi.org/10.14711/thesis-hdl167718","is_oa":false,"landing_page_url":"https://doi.org/10.14711/thesis-hdl167718","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"The Hong Kong University of Science and Technology","raw_type":"dissertation"},"volume":null,"type":"dissertation","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5100618741","display_name":"Ao Sun","orcid":"https://orcid.org/0009-0007-5369-624X","observed_orcids":["https://orcid.org/0009-0007-5369-624X"]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Ao Sun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5100618741"],"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.0,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.0,"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/explainable-ai","display_name":"explainable AI","score":0.9810000061988831},{"id":"https://openalex.org/keywords/concept-based-explanations","display_name":"concept-based explanations","score":0.578000009059906},{"id":"https://openalex.org/keywords/segment-anything-model","display_name":"Segment Anything Model","score":0.8059999942779541},{"id":"https://openalex.org/keywords/large-language-models","display_name":"large language models","score":0.8190000057220459},{"id":"https://openalex.org/keywords/vision-language-models","display_name":"vision-language models","score":0.9919999837875366},{"id":"https://openalex.org/keywords/information-leakage","display_name":"information leakage","score":0.5740000009536743},{"id":"https://openalex.org/keywords/explainability-evaluation","display_name":"explainability evaluation","score":0.3409999907016754}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.609499990940094},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.5787000060081482},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3327000141143799},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.3262999951839447},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.29100000858306885},{"id":"https://openalex.org/C199639397","wikidata":"https://www.wikidata.org/wiki/Q1788588","display_name":"Engineering drawing","level":1,"score":0.27970001101493835},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.2793999910354614},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.273499995470047},{"id":"https://openalex.org/C201995342","wikidata":"https://www.wikidata.org/wiki/Q682496","display_name":"Systems engineering","level":1,"score":0.26080000400543213},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.2565999925136566}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.14711/thesis-hdl167718","doi":"https://doi.org/10.14711/thesis-hdl167718","is_oa":false,"landing_page_url":"https://doi.org/10.14711/thesis-hdl167718","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"The Hong Kong University of Science and Technology","raw_type":"dissertation"},{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-167718","doi":null,"is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-167718","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Thesis"}],"best_oa_location":null,"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W4394973237","https://openalex.org/W7127050500","https://openalex.org/W4400439461","https://openalex.org/W4415318041","https://openalex.org/W6912269051","https://openalex.org/W4414981060","https://openalex.org/W4400975668","https://openalex.org/W4387623017","https://openalex.org/W4401024451","https://openalex.org/W4394597894"],"abstract_inverted_index":{"Recent":[0],"advances":[1],"in":[2,9,21,43,91,195,222],"neural":[3],"networks":[4],"have":[5],"driven":[6],"remarkable":[7],"progress":[8],"computer":[10],"vision":[11],"(CV),":[12],"yet":[13],"their":[14,89,174,283],"opaque":[15],"decision-making":[16],"has":[17],"fueled":[18],"growing":[19],"interest":[20],"Explainable":[22],"AI":[23],"(XAI).":[24],"Conventional":[25],"explanation":[26],"methods":[27],"often":[28],"depend":[29],"on":[30,247],"linear":[31],"segmentation":[32],"and":[33,39,49,82,85,149,180,203,235,281],"manual":[34],"annotations,":[35],"limiting":[36],"faithfulness,":[37],"understandability,":[38],"scalability.":[40],"Meanwhile,":[41],"breakthroughs":[42],"transformer-based":[44,193],"Large":[45],"Language":[46,51],"Models":[47,52],"(LLMs)":[48],"Vision":[50],"(VLMs)":[53],"offer":[54],"new":[55],"opportunities":[56],"to":[57,76,114,139,164,214,259,262,275,285],"produce":[58],"automated,":[59],"high-quality":[60],"concept-explanations.":[61],"In":[62],"this":[63,93,266],"thesis,":[64],"we":[65,86,99,129,199],"investigate":[66],"how":[67],"such":[68],"models":[69,194,274],"can":[70],"serve":[71,276],"as":[72,277,288],"enablers":[73,278],"of":[74,80,153,185,192,279],"XAI":[75],"enhance":[77],"the":[78,101,109,131,182,190,201],"generation":[79],"visual":[81,97],"textual":[83,127],"concepts,":[84,98,128],"further":[87],"assess":[88,189],"limitations":[90,191],"fulfilling":[92],"role":[94],"effectively.":[95],"For":[96,126],"propose":[100,200],"Explain":[102],"Any":[103],"Concept":[104],"(EAC)":[105],"framework,":[106],"which":[107,136,257],"leverages":[108,137],"Segment":[110],"Anything":[111],"Model":[112,134],"(SAM)":[113],"faithfully":[115],"identify":[116],"human-understandable":[117],"image":[118],"regions":[119],"that":[120,144],"influence":[121],"a":[122,170,210,228,260,269,289],"target":[123],"model\u2019s":[124,212],"predictions.":[125],"introduce":[130],"Hierarchical-Concept":[132],"Bottleneck":[133],"(Hi-CBM),":[135],"LLMs":[138],"automatically":[140],"generate":[141,215],"conceptual":[142],"annotations":[143],"are":[145],"both":[146],"richly":[147],"informative":[148],"well-structured.":[150],"The":[151],"richness":[152],"these":[154],"concepts":[155,217],"allows":[156],"any":[157],"CV":[158],"model,":[159],"when":[160],"processing":[161],"an":[162],"image,":[163],"perform":[165],"faithful":[166],"reasoning":[167],"before":[168],"making":[169,230],"final":[171],"decision,":[172],"while":[173],"structured":[175],"organization":[176],"filters":[177],"redundant":[178],"information":[179],"mitigates":[181],"information-leakage":[183],"issues":[184],"traditional":[186],"CBMs.":[187],"To":[188],"generating":[196],"reliable":[197],"explanations,":[198],"Fast":[202],"Slow":[204],"Effect":[205],"(FSE)":[206],"framework.":[207],"FSE":[208],"assesses":[209],"single":[211],"ability":[213],"effective":[216],"by":[218,243,254],"comparing":[219],"classification":[220],"performance":[221],"two":[223],"modes:":[224],"fast":[225,241,252],"mode,":[226],"simulating":[227],"black-box":[229],"direct":[231],"predictions":[232],"without":[233],"rationales,":[234],"purpose":[236],"datasets,":[237,249],"slow":[238],"mode":[239,242,253],"outperforms":[240],"roughly":[244],"10%,":[245],"whereas":[246],"specialized":[248,286],"it":[250],"underperforms":[251],"about":[255],"30%,":[256],"points":[258],"sensitivity":[261],"task":[263],"complexity.":[264],"Overall,":[265],"thesis":[267],"lays":[268],"strong":[270],"foundation":[271],"for":[272,292],"transformer":[273],"XAI,":[280],"highlights":[282],"understandability":[284],"tasks":[287],"critical":[290],"frontier":[291],"future":[293],"research.":[294]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-01-17T00:00:00"}