Matches in SemOpenAlex for { <https://semopenalex.org/work/W3199693760> ?p ?o ?g. }
- W3199693760 endingPage "3089" @default.
- W3199693760 startingPage "3081" @default.
- W3199693760 abstract "Knowledge-based visual question answering (VQA) involves answering questions that require external knowledge not present in the image. Existing methods first retrieve knowledge from external resources, then reason over the selected knowledge, the input image, and question for answer prediction. However, this two-step approach could lead to mismatches that potentially limit the VQA performance. For example, the retrieved knowledge might be noisy and irrelevant to the question, and the re-embedded knowledge features during reasoning might deviate from their original meanings in the knowledge base (KB). To address this challenge, we propose PICa, a simple yet effective method that Prompts GPT3 via the use of Image Captions, for knowledge-based VQA. Inspired by GPT-3’s power in knowledge retrieval and question answering, instead of using structured KBs as in previous work, we treat GPT-3 as an implicit and unstructured KB that can jointly acquire and process relevant knowledge. Specifically, we first convert the image into captions (or tags) that GPT-3 can understand, then adapt GPT-3 to solve the VQA task in a few-shot manner by just providing a few in-context VQA examples. We further boost performance by carefully investigating: (i) what text formats best describe the image content, and (ii) how in-context examples can be better selected and used. PICa unlocks the first use of GPT-3 for multimodal tasks. By using only 16 examples, PICa surpasses the supervised state of the art by an absolute +8.6 points on the OK-VQA dataset. We also benchmark PICa on VQAv2, where PICa also shows a decent few-shot performance." @default.
- W3199693760 created "2021-09-27" @default.
- W3199693760 creator A5021200172 @default.
- W3199693760 creator A5025592561 @default.
- W3199693760 creator A5027851405 @default.
- W3199693760 creator A5048295582 @default.
- W3199693760 creator A5050209478 @default.
- W3199693760 creator A5066666034 @default.
- W3199693760 creator A5073435344 @default.
- W3199693760 date "2022-06-28" @default.
- W3199693760 modified "2023-10-14" @default.
- W3199693760 title "An Empirical Study of GPT-3 for Few-Shot Knowledge-Based VQA" @default.
- W3199693760 cites W1566289585 @default.
- W3199693760 cites W1861492603 @default.
- W3199693760 cites W1933349210 @default.
- W3199693760 cites W2231285690 @default.
- W3199693760 cites W2560730294 @default.
- W3199693760 cites W2886641317 @default.
- W3199693760 cites W2889792105 @default.
- W3199693760 cites W2891394954 @default.
- W3199693760 cites W2947312908 @default.
- W3199693760 cites W2950761309 @default.
- W3199693760 cites W2963115613 @default.
- W3199693760 cites W2963341956 @default.
- W3199693760 cites W2963609017 @default.
- W3199693760 cites W2963717374 @default.
- W3199693760 cites W2963991868 @default.
- W3199693760 cites W2964207259 @default.
- W3199693760 cites W2964303913 @default.
- W3199693760 cites W2965373594 @default.
- W3199693760 cites W2970608575 @default.
- W3199693760 cites W2971869958 @default.
- W3199693760 cites W3016923549 @default.
- W3199693760 cites W3030163527 @default.
- W3199693760 cites W3034972674 @default.
- W3199693760 cites W3093200502 @default.
- W3199693760 cites W3093871960 @default.
- W3199693760 cites W3101703188 @default.
- W3199693760 cites W3122241445 @default.
- W3199693760 cites W3135367836 @default.
- W3199693760 cites W3139224848 @default.
- W3199693760 cites W3172845486 @default.
- W3199693760 cites W3173220247 @default.
- W3199693760 cites W3177174258 @default.
- W3199693760 cites W3177813494 @default.
- W3199693760 doi "https://doi.org/10.1609/aaai.v36i3.20215" @default.
- W3199693760 hasPublicationYear "2022" @default.
- W3199693760 type Work @default.
- W3199693760 sameAs 3199693760 @default.
- W3199693760 citedByCount "20" @default.
- W3199693760 countsByYear W31996937602022 @default.
- W3199693760 countsByYear W31996937602023 @default.
- W3199693760 crossrefType "journal-article" @default.
- W3199693760 hasAuthorship W3199693760A5021200172 @default.
- W3199693760 hasAuthorship W3199693760A5025592561 @default.
- W3199693760 hasAuthorship W3199693760A5027851405 @default.
- W3199693760 hasAuthorship W3199693760A5048295582 @default.
- W3199693760 hasAuthorship W3199693760A5050209478 @default.
- W3199693760 hasAuthorship W3199693760A5066666034 @default.
- W3199693760 hasAuthorship W3199693760A5073435344 @default.
- W3199693760 hasBestOaLocation W31996937601 @default.
- W3199693760 hasConcept C115961682 @default.
- W3199693760 hasConcept C120567893 @default.
- W3199693760 hasConcept C13280743 @default.
- W3199693760 hasConcept C151730666 @default.
- W3199693760 hasConcept C154945302 @default.
- W3199693760 hasConcept C162324750 @default.
- W3199693760 hasConcept C185798385 @default.
- W3199693760 hasConcept C187736073 @default.
- W3199693760 hasConcept C204321447 @default.
- W3199693760 hasConcept C205649164 @default.
- W3199693760 hasConcept C23123220 @default.
- W3199693760 hasConcept C2779343474 @default.
- W3199693760 hasConcept C2780451532 @default.
- W3199693760 hasConcept C41008148 @default.
- W3199693760 hasConcept C44291984 @default.
- W3199693760 hasConcept C4554734 @default.
- W3199693760 hasConcept C86803240 @default.
- W3199693760 hasConceptScore W3199693760C115961682 @default.
- W3199693760 hasConceptScore W3199693760C120567893 @default.
- W3199693760 hasConceptScore W3199693760C13280743 @default.
- W3199693760 hasConceptScore W3199693760C151730666 @default.
- W3199693760 hasConceptScore W3199693760C154945302 @default.
- W3199693760 hasConceptScore W3199693760C162324750 @default.
- W3199693760 hasConceptScore W3199693760C185798385 @default.
- W3199693760 hasConceptScore W3199693760C187736073 @default.
- W3199693760 hasConceptScore W3199693760C204321447 @default.
- W3199693760 hasConceptScore W3199693760C205649164 @default.
- W3199693760 hasConceptScore W3199693760C23123220 @default.
- W3199693760 hasConceptScore W3199693760C2779343474 @default.
- W3199693760 hasConceptScore W3199693760C2780451532 @default.
- W3199693760 hasConceptScore W3199693760C41008148 @default.
- W3199693760 hasConceptScore W3199693760C44291984 @default.
- W3199693760 hasConceptScore W3199693760C4554734 @default.
- W3199693760 hasConceptScore W3199693760C86803240 @default.
- W3199693760 hasIssue "3" @default.
- W3199693760 hasLocation W31996937601 @default.
- W3199693760 hasLocation W31996937602 @default.