LFM-2b-DemoBias

The 'LFM-2b-DemoBias' dataset is used to evaluate the effectiveness of bias mitigation methods in recommendation systems by assessing the ability to obscure protected user attributes, such as gender, while maintaining recommendation performance.

Papers using LFM-2b-DemoBias (1)

Unlearning Protected User Attributes in Recommendations with Adversarial Training2022