Adaptive Autoguidance for Item-Side Fairness in Diffusion Recommender Systems
For recommender system practitioners, this work offers a method to reduce popularity bias in diffusion models, though gains are incremental over existing guided approaches.
A2G-DiffRec addresses popularity bias in diffusion recommenders by adaptively weighting a main model and a weak version of itself, guided by a fairness-aware regularization. It improves item-side fairness with minimal accuracy loss across three datasets.
Diffusion recommender systems achieve strong recommendation accuracy but often suffer from popularity bias, resulting in unequal item exposure. To address this shortcoming, we introduce A2G-DiffRec, a diffusion recommender that incorporates adaptive autoguidance, where the main model is guided by a less-trained version of itself. Instead of using a fixed guidance weight, A2G-DiffRec learns to adaptively weigh the outputs of the main and weak models during training, supervised by a fairness-aware regularization that promotes balanced exposure across items with different popularity levels. Experimental results on three public datasets show that A2G-DiffRec is effective in enhancing item-side fairness at a marginal cost of accuracy reduction compared to existing guided diffusion recommenders and other non-diffusion baselines.