arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.01206cs.AI
收藏

通过SMILE(统计模型无关可解释性与局部解释)解决生成AI的可解释性问题

Addressing Explainability of Generative AI using SMILE (Statistical Model-agnostic Interpretability with Local Explanations)

School of Computer Science(计算机科学学院);University of Hull(赫尔大学)

Zeinab Dehghani

AI总结:

gSMILE通过受控扰动、Wasserstein距离和加权替代模型,提升生成AI的可解释性,实现细粒度归因和直观热图生成。

摘要:

生成人工智能的快速发展使模型能够生成复杂的文本和视觉输出;然而,其决策过程仍然大多模糊,限制了在高风险应用中的信任和问责。本论文介绍了一个统一的框架gSMILE,用于生成模型的可解释性,将统计模型无关可解释性与局部解释(SMILE)方法扩展到生成领域。gSMILE通过受控的文本输入扰动、Wasserstein距离度量和加权替代模型来量化和可视化提示或指令中特定组件如何影响模型输出。应用于大型语言模型(LLMs)时,gSMILE提供细粒度的令牌级归因,并生成直观的热图,突出显示有影响力的令牌和推理路径。在基于指令的图像编辑模型中,精确的文本扰动机制被使用,允许分析对编辑指令的修改如何影响最终图像。结合基于场景的评估策略,该策略基于操作设计领域(ODD)框架,gSMILE允许在多样化的语义和环境条件下系统评估模型行为。为了评估解释质量,我们定义了严格的归因指标,包括稳定性、保真度、准确性、一致性、忠实度,并将其应用于多种生成架构。广泛的实验表明,gSMILE生成的归因具有鲁棒性、与人类一致,并且在最先进生成模型上具有良好的泛化能力。这些发现突显了gSMILE在推动透明、可靠和负责任的生成AI技术部署方面的潜力。

英文摘要:

The rapid advancement of generative artificial intelligence has enabled models capable of producing complex textual and visual outputs; however, their decision-making processes remain largely opaque, limiting trust and accountability in high-stakes applications. This thesis introduces gSMILE, a unified framework for the explainability of generative models, extending the Statistical Model-agnostic Interpretability with Local Explanations (SMILE) method to generative settings. gSMILE employs controlled perturbations of textual input, Wasserstein distance metrics, and weighted surrogate modelling to quantify and visualise how specific components of a prompt or instruction influence model outputs. Applied to Large Language Models (LLMs), gSMILE provides fine-grained token-level attribution and generates intuitive heatmaps that highlight influential tokens and reasoning pathways. In instruction-based image editing models, the exact text-perturbation mechanism is employed, allowing for the analysis of how modifications to an editing instruction impact the resulting image. Combined with a scenario-based evaluation strategy grounded in the Operational Design Domain (ODD) framework, gSMILE allows systematic assessment of model behaviour across diverse semantic and environmental conditions. To evaluate explanation quality, we define rigorous attribution metrics, including stability, fidelity, accuracy, consistency, and faithfulness, and apply them across multiple generative architectures. Extensive experiments demonstrate that gSMILE produces robust, human-aligned attributions and generalises effectively across state-of-the-art generative models. These findings highlight the potential of gSMILE to advance transparent, reliable, and responsible deployment of generative AI technologies.

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。

↑