AOEPT proposes modal-contextualized prompts that distill global modality priors to restore reasoning scope in multimodal transformers under missing-modality conditions.
Prompting through prototype: A prototype-based prompt learning on pretrained vision-language models.arXiv preprint arXiv:2210.10841,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
AOEPT: Breaking the Implicit Modality-Reduction Bottleneck in Modality-Missing Prompt Tuning
AOEPT proposes modal-contextualized prompts that distill global modality priors to restore reasoning scope in multimodal transformers under missing-modality conditions.