Fairness-Aware Test-Time Prompt Tuning
cs.LG
Submitted: 2026-08-26
Updated: 2026-08-26
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
The gist: Vision-language models have displayed remarkable capabilities in multi-modal understanding and are increasingly used in critical applications where economic and practical deployment constraints
Terminology
Abstract
Vision-language models have displayed remarkable capabilities in multi-modal understanding and are increasingly used in critical applications where economic and practical deployment constraints prohibit re-training or fine-tuning. However, these models can also exhibit systematic biases that disproportionately affect protected demographic groups and existing approaches to addressing these biases require extensive model retraining and access to demographic attributes. There is a clear need to develop test-time adaptation (TTA) approaches that improve the fairness characteristics of pretrained models under distributional shift. In this paper, we evaluate how episodic TTA affects fairness in CLIP classification under subpopulation shifts and develop FairTPT, a novel fairness-aware episodic TTA method that jointly minimizes target marginal entropy while maximizing spurious marginal entropy through soft-prompt tuning. We find that standard episodic TTA generally exacerbates disparities between majority and minority groups, that blinding a model to spurious attributes without degrading target performance is inherently challenging, and that excessive blinding can lead to catastrophic forgetting. This model collapse can be prevented by monitoring test-time changes in target loss within the linear regime, while still achieving fairness improvements on reactive data and preserving overall performance. FairTPT outperforms all state-of-the-art episodic test-time debiasing methods and establishes a foundation for robust TTA, which is essential for achieving fairness in practice.
Sources
- Multimodal datasets: misogyny, pornography, and malignant stereotypes
- Evaluating CLIP: Towards Characterization of Broader Capabilities and Downstream Implications
- Identifying Implicit Social Biases in Vision-Language Models
- Vision-Language Models display a strong gender bias
- Vision-Language Models Performing Zero-Shot Tasks Exhibit Gender-based Disparities
- Balancing Discriminability and Transferability for Source-Free Domain Adaptation
- The Illusion of Progress? A Critical Look at Test-Time Adaptation for Vision-Language Models
- Debiasing Vision-Language Models via Biased Prompts
- Beyond Model Adaptation at Test Time: A Survey
- Invariant Risk Minimization
- PerceptionCLIP: Visual Classification by Inferring and Conditioning on Contexts
- Jacobian Descent for Multi-Objective Optimization
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks