Deeper Diffusion Models Amplify Bias
It addresses bias amplification and privacy risks in generative AI models, which is an incremental but important safety concern for AI practitioners.
This paper investigates how deeper diffusion models can amplify inherent biases in training data while potentially compromising privacy, with theoretical and empirical validation showing these effects increase with model depth.
Despite the remarkable performance of generative Diffusion Models (DMs), their internal working is still not well understood, which is potentially problematic. This paper focuses on exploring the important notion of bias-variance tradeoff in diffusion models. Providing a systematic foundation for this exploration, it establishes that at one extreme, the diffusion models may amplify the inherent bias in the training data, and on the other, they may compromise the presumed privacy of the training samples. Our exploration aligns with the memorization-generalization understanding of the generative models, but it also expands further along this spectrum beyond "generalization", revealing the risk of bias amplification in deeper models. Our claims are validated both theoretically and empirically.