Contribution-Aware Fair Recommendation
What it is
This paper proposes Contribution-Performance Fairness, a new perspective for recommender systems that aligns user recommendation benefits with their estimated contributions to model learning. It introduces the Contribution-Performance Fair Recommender (CPFR) framework, which constructs user groups based on interaction volume, loss alignment, and optimization intensity to jointly optimize accuracy and fairness.
The problem it solves
Existing fairness research in recommender systems ignores the distributive perspective of whether users' contributions to model learning are reflected in the benefits they receive.
What you could use it for
- Designing recommender systems that incentivize sustained user engagement by rewarding informative contributions
- Implementing fairness constraints that account for user effort in model training alongside traditional demographic fairness
- Improving system-level recommendation accuracy in voluntary contribution ecosystems
Read the source ↗Back to the library →
Summary written by an AI model from the source; check the source before you rely on it.