Olivia Wiles

Revisiting Text-to-Image Evaluation with Gecko: On Metrics, Prompts, and Human Ratings

Generative models improve fairness of medical classifiers under distribution shifts

Evaluating model bias requires characterizing model mistakes

Adversarially self-supervised pre-training improves accuracy and robustness

Differentially private diffusion models generate useful synthetic images

Discovering bugs in vision models using off-the-shelf image generation and captioning

Compressed vision for efficient video understanding

Data augmentation can improve robustness