2023最新CV 推荐论文及代码合集Prompting Large Language Models with Answer Heuristics for__Knowledge-based Visual Question Answering.pdfVIP
- 0
- 0
- 约16.05万字
- 约 16页
- 2026-01-28 发布于浙江
- 举报
PromptingLargeLanguageModelswithAnswerHeuristicsfor
Knowledge-basedVisualQuestionAnswering
ZhenweiShao1ZhouYu1*MengWang2JunYu1
1KeyLaboratoryofComplexSystemsModelingandSimulation,
SchoolofComputerScienceandTechnology,HangzhouDianziUniversity,China.
2SchoolofComputerScienceandInformationEngineering,HefeiUniversityofTechnology,China
3
2shaozw,yuz,yujun@,eric.mengwang@
0Code:/MILVLG/prophet
2
r
a
whatfruitcomesPICa
MAbstract(Q)fromthesetrees?CGPT-3A
entity-1,desc-1
(K)...Q
3Knowledge-basedvisualquestionanswering(VQA)re-entity-N,desc-N
KAT/REVIVE
]quiresexternalknowledgebeyondtheimagetoanswertheknowledgebaseC
candidatesKB-augmented
Vquestion.EarlystudiesretrieverequiredknowledgefromGPTGPT--33evidenceVQAmodelA
原创力文档

文档评论(0)