GeNIe: Generative Hard Negative Images Through Diffusion
This work addresses the need for more effective data augmentation to prevent overfitting in deep learning, particularly in scenarios with limited or imbalanced data, representing an incremental improvement over existing generative augmentation techniques.
The paper tackles the problem of data augmentation for deep models by introducing GeNIe, a method that uses a latent diffusion model to generate hard negative images by combining source images with target text prompts, achieving superior performance over prior methods in few-shot and long-tail settings.
Data augmentation is crucial in training deep models, preventing them from overfitting to limited data. Recent advances in generative AI, e.g., diffusion models, have enabled more sophisticated augmentation techniques that produce data resembling natural images. We introduce GeNIe a novel augmentation method which leverages a latent diffusion model conditioned on a text prompt to combine two contrasting data points (an image from the source category and a text prompt from the target category) to generate challenging augmentations. To achieve this, we adjust the noise level (equivalently, number of diffusion iterations) to ensure the generated image retains low-level and background features from the source image while representing the target category, resulting in a hard negative sample for the source category. We further automate and enhance GeNIe by adaptively adjusting the noise level selection on a per image basis (coined as GeNIe-Ada), leading to further performance improvements. Our extensive experiments, in both few-shot and long-tail distribution settings, demonstrate the effectiveness of our novel augmentation method and its superior performance over the prior art. Our code is available at: https://github.com/UCDvision/GeNIe