Summary of Adaptive Fairness-aware Online Meta-learning For Changing Environments, by Chen Zhao et al.
Adaptive Fairness-Aware Online Meta-Learning for Changing Environments
by Chen Zhao, Feng Mi, Xintao Wu, Kai Jiang, Latifur Khan, Feng Chen
First submitted to arxiv on: 20 May 2022
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed fair-aware online learning framework enables learners to sequentially acquire new tasks while ensuring statistical parity across protected sub-populations. Existing methods rely heavily on i.i.d data assumptions, providing static regret analysis, but this doesn’t guarantee good performance in changing environments with heterogeneous task distributions. To address this issue, the authors introduce a novel regret metric FairSAR by incorporating long-term fairness constraints into a strongly adapted loss regret. They also propose the FairSAOML algorithm, which adapts to changing environments in both bias control and model precision. The framework is formulated as a bi-level convex-concave optimization problem, providing sub-linear upper bounds for loss regret and cumulative fairness constraint violation. Experimental results on real-world datasets demonstrate that FairSAOML outperforms prior online learning approaches. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper introduces an online learning system that helps people learn new skills over time while making sure they’re fair to everyone. Right now, most systems assume that the data is the same every time, but this isn’t always true. The authors create a new way to measure how well the system does and propose a new algorithm to adapt to changing situations. They test their approach on real-world datasets and show it performs better than existing methods. |
Keywords
* Artificial intelligence * Online learning * Optimization * Precision
