BalancedDPO: Adaptive Multi-Metric Alignment
This addresses the problem of limited generalization in text-to-image models for users needing robust multi-metric alignment, though it is incremental as it extends an existing method.
The paper tackles the challenge of aligning text-to-image diffusion models with diverse preferences by introducing BalancedDPO, a method that simultaneously aligns models with multiple metrics like human preference, CLIP score, and aesthetic quality, achieving state-of-the-art results with average win rate improvements of 15%, 7.1%, and 10.3% on benchmark datasets.
Text-to-image (T2I) diffusion models have made remarkable advancements, yet aligning them with diverse preferences remains a persistent challenge. Current methods often optimize single metrics or depend on narrowly curated datasets, leading to overfitting and limited generalization across key visual quality metrics. We present BalancedDPO, a novel extension of Direct Preference Optimization (DPO) that addresses these limitations by simultaneously aligning T2I diffusion models with multiple metrics, including human preference, CLIP score, and aesthetic quality. Our key novelty lies in aggregating consensus labels from diverse metrics in the preference distribution space as compared to existing reward mixing approaches, enabling robust and scalable multi-metric alignment while maintaining the simplicity of the standard DPO pipeline that we refer to as BalancedDPO. Our evaluations on the Pick-a-Pic, PartiPrompt and HPD datasets show that BalancedDPO achieves state-of-the-art results, outperforming existing approaches across all major metrics. BalancedDPO improves the average win rates by 15%, 7.1%, and 10.3% on Pick-a-pic, PartiPrompt and HPD, respectively, from the DiffusionDPO.