MetaVL: Transferring In-Context Learning Ability From Language Models to Vision-Language Models
This addresses the lack of in-context learning in vision-language models, which is incremental as it adapts existing language model techniques to a multimodal domain.
The paper tackles enabling in-context learning for vision-language models by transferring this ability from language models, resulting in a model that outperforms baselines on VQA, OK-VQA, and GQA with 20 times fewer parameters.
Large-scale language models have shown the ability to adapt to a new task via conditioning on a few demonstrations (i.e., in-context learning). However, in the vision-language domain, most large-scale pre-trained vision-language (VL) models do not possess the ability to conduct in-context learning. How can we enable in-context learning for VL models? In this paper, we study an interesting hypothesis: can we transfer the in-context learning ability from the language domain to VL domain? Specifically, we first meta-trains a language model to perform in-context learning on NLP tasks (as in MetaICL); then we transfer this model to perform VL tasks by attaching a visual encoder. Our experiments suggest that indeed in-context learning ability can be transferred cross modalities: our model considerably improves the in-context learning capability on VL tasks and can even compensate for the size of the model significantly. On VQA, OK-VQA, and GQA, our method could outperform the baseline model while having 20 times fewer parameters.