Imbalance-Robust and Sampling-Efficient Continuous Conditional GANs via Adaptive Vicinity and Auxiliary Regularization
This addresses data imbalance and sampling inefficiency in conditional generative modeling for applications like image generation with continuous labels, though it is incremental as it builds on existing CcGAN frameworks.
The paper tackles the limitations of continuous conditional generative models, specifically data imbalance in CcGAN and slow sampling in CCDM, by proposing CcGAN-AVAR, which achieves state-of-the-art generation quality with 30x-2000x faster inference than CCDM.
Recent advances in conditional generative modeling have introduced Continuous conditional Generative Adversarial Network (CcGAN) and Continuous Conditional Diffusion Model (CCDM) for estimating high-dimensional data distributions conditioned on scalar, continuous regression labels (e.g., angles, ages, or temperatures). However, these approaches face fundamental limitations: CcGAN suffers from data imbalance due to fixed-size vicinity constraints, while CCDM requires computationally expensive iterative sampling. To address these issues, we propose CcGAN-AVAR, an enhanced CcGAN framework featuring (1) two novel components for handling data imbalance - an adaptive vicinity mechanism that dynamically adjusts vicinity size and a multi-task discriminator that enhances generator training through auxiliary regression and density ratio estimation - and (2) the GAN framework's native one-step generator, enable 30x-2000x faster inference than CCDM. Extensive experiments on four benchmark datasets (64x64 to 256x256 resolution) across eleven challenging settings demonstrate that CcGAN-AVAR achieves state-of-the-art generation quality while maintaining sampling efficiency.