Computer vision models have known performance disparities across attribu...
The unprecedented photorealistic results achieved by recent text-to-imag...
Disaggregated performance metrics across demographic groups are a hallma...
We explore the extent to which zero-shot vision-language models exhibit
...
Unwanted and often harmful social biases are becoming ever more salient ...
We present a novel task and dataset for evaluating the ability of vision...
We introduce CM3, a family of causally masked generative models trained ...
Children acquire their native language with apparent ease by observing h...
We generalize the notion of social biases from language embeddings to
gr...