We study the proportional clustering problem of Chen et al. (ICML '19) and relate

Toaddress this,we introduce anovelfairrecommendation framework,Problem(FAIR),thatflexibly balances multi-stakeholder interests via a constrained optimization formulation.

We used the sex and the education of the student's parents as the sensitive attributes for this dataset. We removed all features that are other expressions of the labels (i.e. Note that this is the only folktables dataset on which we report results in the main paper. Sex, age, and rage are used as sensitive features for this datasets. We deem these features as not relevant for this use case.

Wesummarize all unidentifiable situations that are discovered in the causal inference literature. Then, we develop a constrained optimization problem forbounding thePCfairness, whichismotivatedbythemethod proposed in[2]forbounding confounded causaleffects. Thekeyideaistoparameterize thecausal model using so-called response-function variables, whose distribution captures all randomness encoded in the causal model, so that we can explicitly traverse all possible causal models to find thetightest possible bounds.

In this work, we mainly focus onthe notions ofgroup fairness.

Fairness is a central pillar of trustworthy machine learning, especially in domains where accuracy- or profit-driven optimization is insufficient. While most fairness research focuses on supervised learning, fairness in policy learning remains less explored. Because policy learning is interventional, it induces two distinct fairness targets: action fairness (equitable action assignments) and outcome fairness (equitable downstream consequences). Crucially, equalizing actions does not generally equalize outcomes when groups face different constraints or respond differently to the same action. We propose a novel double fairness learning (DFL) framework that explicitly manages the trade-off among three objectives: action fairness, outcome fairness, and value maximization. We integrate fairness directly into a multi-objective optimization problem for policy learning and employ a lexicographic weighted Tchebyshev method that recovers Pareto solutions beyond convex settings, with theoretical guarantees on the regret bounds. Our framework is flexible and accommodates various commonly used fairness notions. Extensive simulations demonstrate improved performance relative to competing methods. In applications to a motor third-party liability insurance dataset and an entrepreneurship training dataset, DFL substantially improves both action and outcome fairness while incurring only a modest reduction in overall value.