Balancing Fairness and Robustness via Partial Invariance
This work addresses a failure case in OOD generalization for networked or hierarchical data, offering a method to improve fairness and robustness, though it appears incremental as it builds on the existing IRM framework.
The paper tackles the problem of suboptimal performance in Invariant Risk Minimization (IRM) when data overlap assumptions fail, by proposing a partial invariance framework that partitions environments based on hierarchical differences and enforces invariance locally. The results show this approach can alleviate the trade-off between fairness and risk in classification settings with causal distribution shifts.
The Invariant Risk Minimization (IRM) framework aims to learn invariant features from a set of environments for solving the out-of-distribution (OOD) generalization problem. The underlying assumption is that the causal components of the data generating distributions remain constant across the environments or alternately, the data "overlaps" across environments to find meaningful invariant features. Consequently, when the "overlap" assumption does not hold, the set of truly invariant features may not be sufficient for optimal prediction performance. Such cases arise naturally in networked settings and hierarchical data-generating models, wherein the IRM performance becomes suboptimal. To mitigate this failure case, we argue for a partial invariance framework. The key idea is to introduce flexibility into the IRM framework by partitioning the environments based on hierarchical differences, while enforcing invariance locally within the partitions. We motivate this framework in classification settings with causal distribution shifts across environments. Our results show the capability of the partial invariant risk minimization to alleviate the trade-off between fairness and risk in certain settings.