Bias Begets Bias: The Impact of Biased Embeddings on Diffusion Models
Published in ICML 2024 Workshop on Trustworthy Multi-modal Foundation Models and AI Agents, 2024
We investigate how biases in text/image embeddings amplify downstream harm when used to train diffusion models.
Recommended citation: Sahil Kuchlous*, Marvin Li*, and Jeffrey G. Wang*. (2024). "Bias Begets Bias: The Impact of Biased Embeddings on Diffusion Models." ICML Workshop on Trustworthy Multi-modal Foundation Models and AI Agents.
Download Paper