PRISM: Reducing Spurious Implicit Biases in Vision-Language Models with LLM-Guided Embedding Projection
This addresses bias mitigation in VLMs, a critical issue for fairness in AI applications, though it appears incremental as it builds on existing debiasing approaches with a novel projection technique.
The paper tackles the problem of spurious biases in vision-language models (VLMs) like CLIP, which can lead to skewed predictions, by introducing PRISM, a data-free and task-agnostic method that uses LLM-guided embedding projection to reduce these biases, achieving superior performance on Waterbirds and CelebA datasets compared to existing debiasing methods.
We introduce Projection-based Reduction of Implicit Spurious bias in vision-language Models (PRISM), a new data-free and task-agnostic solution for bias mitigation in VLMs like CLIP. VLMs often inherit and amplify biases in their training data, leading to skewed predictions. PRISM is designed to debias VLMs without relying on predefined bias categories or additional external data. It operates in two stages: first, an LLM is prompted with simple class prompts to generate scene descriptions that contain spurious correlations. Next, PRISM uses our novel contrastive-style debiasing loss to learn a projection that maps the embeddings onto a latent space that minimizes spurious correlations while preserving the alignment between image and text embeddings.Extensive experiments demonstrate that PRISM outperforms current debiasing methods on the commonly used Waterbirds and CelebA datasets We make our code public at: https://github.com/MahdiyarMM/PRISM.