跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.22889 · v1

Beyond Attribute-Centric Fairness: Modelling and Mitigating Unified Bias Representation in Omni-modal Large Language Models

首次提交 2026-11-17 22:59
版本更新 2026-11-17 22:59
雷达观测 2026-11-18 16:38
arXiv ↗PDF ↗

摘要

Omni-modal large language models (OLLMs) provide joint multimodal understanding and generation capabilities, making them well-suited to complex environments and promising backbones for next-generation AI. Despite these advances, the harmful stereotypes encoded in such models manifest in increasingly diverse forms, leading to broader and more severe societal consequences. Existing debiasing methods typically impose separate fairness objectives for each sensitive attribute and modality to minimise demographic-dependent variations. However, in omni-modal settings, this attribute-centric paradigm struggles to jointly address diverse social biases with distinct fairness objectives and to generalise beyond predefined attributes. To this end, we propose OmniDebias, one of the first methods for promoting fairness in OLLMs, which employs a bias-centric formulation to extract and mitigate a shared bias representation across attributes and modalities while preserving attribute-related knowledge. Specifically, we exploit activation inconsistencies of a given attribute (e.g., gender) across semantic contexts (e.g., occupations) to disentangle bias from legitimate knowledge. We further consolidate multiple extracted biases into a shared representation, formulating a unified objective that captures the underlying structure across social biases and generalises to unseen attributes. Extensive experiments demonstrate that eliminating unified bias with both post-hoc editing and fine-tuning can consistently improve fairness across diverse attributes and modalities in OLLMs while maintaining favourable utility.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-17 22:59 · 最新观测版本

Beyond Attribute-Centric Fairness: Modelling and Mitigating Unified Bias Representation in Omni-modal Large Language Models

Omni-modal large language models (OLLMs) provide joint multimodal understanding and generation capabilities, making them well-suited to complex environments and promising backbones for next-generation AI. Despite these advances, the harmful stereotypes encoded in such models manifest in increasingly diverse forms, leading to broader and more severe societal consequences. Existing debiasing methods typically impose separate fairness objectives for each sensitive attribute and modality to minimise demographic-dependent variations. However, in omni-modal settings, this attribute-centric paradigm struggles to jointly address diverse social biases with distinct fairness objectives and to generalise beyond predefined attributes. To this end, we propose OmniDebias, one of the first methods for promoting fairness in OLLMs, which employs a bias-centric formulation to extract and mitigate a shared bias representation across attributes and modalities while preserving attribute-related knowledge. Specifically, we exploit activation inconsistencies of a given attribute (e.g., gender) across semantic contexts (e.g., occupations) to disentangle bias from legitimate knowledge. We further consolidate multiple extracted biases into a shared representation, formulating a unified objective that captures the underlying structure across social biases and generalises to unseen attributes. Extensive experiments demonstrate that eliminating unified bias with both post-hoc editing and fine-tuning can consistently improve fairness across diverse attributes and modalities in OLLMs while maintaining favourable utility.

查看此版本解读 →查看此版本原文 ↗