BendVLM: Test-Time Debiasing of Vision-Language Embeddings
Preprint
Abstract
Vision-language model (VLM) embeddings have been shown to encode biases present in their training data, such as societal biases that prescribe negative characteristics to members of various racial and gender identities. VLMs are being quickly adopted for a variety of tasks ranging from few-shot classification to text-guided image generation, making debiasing VLM embeddings crucial. Debiasing approaches that fine-tune the VLM often suffer from catastrophic forgetting. On the other hand, fine-tuning-free methods typically utilize a "one-size-fits-all" approach that assumes that correlation with the spurious attribute can be explained using a single linear direction across all possible inputs. In this work, we propose Bend-VLM, a nonlinear, fine-tuning-free approach for VLM embedding debiasing that tailors the debiasing operation to each unique input. This allows for a more flexible debiasing approach. Additionally, we do not require knowledge of the set of inputs a priori to inference time, making our method more appropriate for online, open-set tasks such as retrieval and text guided image generation.Keywords
All Related Versions
This publication has 8 references indexed in Scilit:
- Improved Baselines with Visual Instruction Tuning2024
- FairCLIP: Harnessing Fairness in Vision-Language Learning2024
- Discovering and Mitigating Visual Biases Through Keyword Explanation2024
- Worst of Both Worlds: Biases Compound in Pre-trained Vision-and-Language Models2022
- Implicit Stereotypes in Pre-Trained ClassifiersIEEE Access, 2021
- Are Gender-Neutral Queries Really Gender-Neutral? Mitigating Gender Bias in Image Search2021
- FairFace: Face Attribute Dataset for Balanced Race, Gender, and Age for Bias Measurement and Mitigation2021
- SciPy 1.0: fundamental algorithms for scientific computing in PythonNature Methods, 2020