arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2312.14895cs.LGcs.AI
收藏

FAST:面向黑盒生成模型弱遗忘的特征感知相似性阈值法

FAST: Feature Aware Similarity Thresholding for Weak Unlearning in Black-Box Generative Models

Indian Institute of Science(印度科学学院)

Subhodip Panda, Prathosh AP

AI总结:

针对黑盒生成模型无法通过传统机器遗忘消除不良输出的问题,本研究提出FAST方法,通过隐空间编码不良特征表征抑制不良输出,并证明过滤属于弱遗忘形式。

摘要:

随着隐私保护及监管合规相关担忧日益加剧,人们对深度生成模型监管的重视程度不断提升,这凸显了为这类模型构建精准控制机制的迫切需求。当生成模型产出包含不良、冒犯性或潜在有害内容的输出时,这种紧迫性尤为突出。为此,机器遗忘技术应运而生,旨在从预训练模型中选择性遗忘特定知识,或消除不良数据子集带来的影响。然而,现有主流机器遗忘方法通常要求在遗忘过程中可获取模型参数与架构细节,这一条件并非总能满足。在众多下游任务中,这类模型以黑盒系统的形式运行,其预训练参数、架构及训练数据均无法获取。在这类场景下,对不良输出进行过滤成为一种可行的替代方案。本研究的核心目标有两个:一是阐明过滤与遗忘过程之间的关联,二是提出一种方法,以减少被归类为黑盒系统的模型所生成的不良输出的展示。本研究的理论分析表明,在黑盒模型场景下,过滤可被视为一种弱遗忘形式。我们提出的**特征感知相似性阈值法(Feature Aware Similarity Thresholding, FAST)**通过在隐空间中对不良特征的表征进行系统性编码,可有效抑制不良输出。

英文摘要:

The heightened emphasis on the regulation of deep generative models, propelled by escalating concerns pertaining to privacy and compliance with regulatory frameworks, underscores the imperative need for precise control mechanisms over these models. This urgency is particularly underscored by instances in which generative models generate outputs that encompass objectionable, offensive, or potentially injurious content. In response, machine unlearning has emerged to selectively forget specific knowledge or remove the influence of undesirable data subsets from pre-trained models. However, modern machine unlearning approaches typically assume access to model parameters and architectural details during unlearning, which is not always feasible. In multitude of downstream tasks, these models function as black-box systems, with inaccessible pre-trained parameters, architectures, and training data. In such scenarios, the possibility of filtering undesired outputs becomes a practical alternative. The primary goal of this study is twofold: first, to elucidate the relationship between filtering and unlearning processes, and second, to formulate a methodology aimed at mitigating the display of undesirable outputs generated from models characterized as black-box systems. Theoretical analysis in this study demonstrates that, in the context of black-box models, filtering can be seen as a form of weak unlearning. Our proposed \textbf{\textit{Feature Aware Similarity Thresholding(FAST)}} method effectively suppresses undesired outputs by systematically encoding the representation of unwanted features in the latent space.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑