Model Adaptation · Software component
Representation Rebalancer
Software componentModel AdaptationModelsarc:RepresentationRebalancer
A pre-processing bias mitigator that balances demographic representation in training data by resampling, instance reweighting, or targeted augmentation of underrepresented groups.
Responsibility. Produces a demographically balanced (or reweighted) training dataset.
Also known as: Fairness-aware data preparation, Data balancing, Pre-processing debiasing
Variant of Bias Mitigator abstract
When to choose. Choose (pre-processing) when training data underrepresents groups or encodes historical imbalance; reweighting preserves the original distribution, while resampling/augmentation change the dataset.
Relationships
invokes dependency
reads dependency
receives data from dynamic
produces lifecycle
alternative to variability
Design guidance
- SHOULD NOT oversample by simple duplication; use synthetic minority examples to avoid overfitting.
- SHOULD target augmentation at underrepresented groups along protected attributes.
Quantitative guidance
As stated by the sources; verify before use.
- Example: 80%/20% male/female resume data resampled toward 50/50 (Ch9.4).
- Example: 1000 majority vs. 200 minority examples receive 5x minority weight (Ch9.4).
Classification
- Patterns
- Undersampling / oversamplingSMOTEReweighing (instance reweighting)Fairness-targeted data augmentationCounterfactual data generation
- Risks mitigated
- Representation biasOverfitting from naive duplication of minority examples
Sources
- Ch9.4: T. Nguyen, "Fairness and Bias Mitigation," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 9.4. ISBN: 9798244538229.
- Ref9.02: "Responsible AI and Ethical Principles," unpublished reference note (02-Responsible-AI-Ethical-Principles.md), Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam supplementary materials, 2026. unpublished note