Mitigating Bias in RAG: Controlling the Embedder
This addresses bias mitigation in RAG systems for AI fairness, though it is incremental as it builds on existing debiasing methods.
The paper tackled bias in retrieval augmented generation (RAG) systems by studying bias conflict among components, showing that reverse-biasing the embedder can mitigate overall bias while maintaining utility, as demonstrated through experiments with 120 differently biased embedders and 6 LLMs.
In retrieval augmented generation (RAG) systems, each individual component -- the LLM, embedder, and corpus -- could introduce biases in the form of skews towards outputting certain perspectives or identities. In this work, we study the conflict between biases of each component and their relationship to the overall bias of the RAG system, which we call bias conflict. Examining both gender and political biases as case studies, we show that bias conflict can be characterized through a linear relationship among components despite its complexity in 6 different LLMs. Through comprehensive fine-tuning experiments creating 120 differently biased embedders, we demonstrate how to control bias while maintaining utility and reveal the importance of reverse-biasing the embedder to mitigate bias in the overall system. Additionally, we find that LLMs and tasks exhibit varying sensitivities to the embedder bias, a crucial factor to consider for debiasing. Our results underscore that a fair RAG system can be better achieved by carefully controlling the bias of the embedder rather than increasing its fairness.