Dynamic Environment Responsive Online Meta-Learning with Fairness Awareness
Chen Zhao, Feng Mi, Xintao Wu, Kai Jiang, Latifur Khan, Feng Chen
TL;DR
This work tackles fairness-aware online meta-learning in non-stationary environments by introducing FairSAR, a fairness-aware strongly adaptive regret, and FairSAOML, a bi-level adaptive online meta-learning algorithm. It frames the problem as a constrained bi-level convex-concave optimization with primal parameters for accuracy and dual parameters for fairness, and handles non-stationarity via three interval-based expert schemes: Dynamic Intervals (DI), Adaptive Geometric Covering (AGC), and Dynamic Geometric Covering (DGC). Theoretical analysis yields sub-linear loss regret $\mathcal{O}((\tau\log T)^{1/2})$ and fairness-violation bounds $\mathcal{O}((\tau T\log T)^{1/4})$, while experiments on NYSF and MovieLens-based datasets show FairSAOML achieving superior accuracy and fairness across changing environments. The approach offers practical impact for deploying fair, adaptive models in real-world, evolving data streams, with efficient, interval-informed expert tracking and a robust meta-learning backbone.
Abstract
The fairness-aware online learning framework has emerged as a potent tool within the context of continuous lifelong learning. In this scenario, the learner's objective is to progressively acquire new tasks as they arrive over time, while also guaranteeing statistical parity among various protected sub-populations, such as race and gender, when it comes to the newly introduced tasks. A significant limitation of current approaches lies in their heavy reliance on the i.i.d (independent and identically distributed) assumption concerning data, leading to a static regret analysis of the framework. Nevertheless, it's crucial to note that achieving low static regret does not necessarily translate to strong performance in dynamic environments characterized by tasks sampled from diverse distributions. In this paper, to tackle the fairness-aware online learning challenge in evolving settings, we introduce a unique regret measure, FairSAR, by incorporating long-term fairness constraints into a strongly adapted loss regret framework. Moreover, to determine an optimal model parameter at each time step, we introduce an innovative adaptive fairness-aware online meta-learning algorithm, referred to as FairSAOML. This algorithm possesses the ability to adjust to dynamic environments by effectively managing bias control and model accuracy. The problem is framed as a bi-level convex-concave optimization, considering both the model's primal and dual parameters, which pertain to its accuracy and fairness attributes, respectively. Theoretical analysis yields sub-linear upper bounds for both loss regret and the cumulative violation of fairness constraints. Our experimental evaluation on various real-world datasets in dynamic environments demonstrates that our proposed FairSAOML algorithm consistently outperforms alternative approaches rooted in the most advanced prior online learning methods.