Z-LaVI: Zero-Shot Language Solver Fueled by Visual Imagination

21 Oct 2022  ·  Yue Yang, Wenlin Yao, Hongming Zhang, Xiaoyang Wang, Dong Yu, Jianshu Chen ·

Large-scale pretrained language models have made significant advances in solving downstream language understanding tasks. However, they generally suffer from reporting bias, the phenomenon describing the lack of explicit commonsense knowledge in written text, e.g., ''an orange is orange''. To overcome this limitation, we develop a novel approach, Z-LaVI, to endow language models with visual imagination capabilities. Specifically, we leverage two complementary types of ''imaginations'': (i) recalling existing images through retrieval and (ii) synthesizing nonexistent images via text-to-image generation. Jointly exploiting the language inputs and the imagination, a pretrained vision-language model (e.g., CLIP) eventually composes a zero-shot solution to the original language tasks. Notably, fueling language models with imagination can effectively leverage visual knowledge to solve plain language tasks. In consequence, Z-LaVI consistently improves the zero-shot performance of existing language models across a diverse set of language tasks.

PDF Abstract
Task Dataset Model Metric Name Metric Value Global Rank Benchmark
Visual Commonsense Tests ViComTe-color Z-LaVI (GPT-J-6B) Spearman's Rho 49.6 # 2


No methods listed for this paper. Add relevant methods here