可解释的人工智能：面向Transformer模型的上下文感知分层集成梯度方法
Explainable AI: Context-Aware Layer-Wise Integrated Gradients for Explaining Transformer Models
AI总结：
本文提出了一种上下文感知分层集成梯度框架（CA-LIG），用于解释Transformer模型的决策过程，通过计算每个Transformer块内的分层集成梯度，并将这些token级属性与类特定的注意力梯度融合，从而生成具有符号和上下文敏感性的属性图，以捕捉支持和反对的证据，并追踪Transformer层中的相关性层次流动。
摘要：
Transformer模型在多个领域和任务中实现了最先进的性能，然而其深层表示使得预测难以解释。现有的可解释性方法依赖于最终层的属性，只能捕捉局部token级属性或全局注意力模式，缺乏对token间依赖关系和结构组件的上下文感知能力。它们还无法捕捉相关性如何在层之间演变以及结构组件如何影响决策。为了解决这些限制，我们提出了上下文感知分层集成梯度（CA-LIG）框架，一种统一的层次属性框架，该框架在每个Transformer块内计算分层集成梯度，并将这些token级属性与类特定的注意力梯度融合。这种整合产生了带有符号和上下文敏感性的属性图，能够捕捉支持和反对的证据，同时追踪Transformer层中的相关性层次流动。我们评估了CA-LIG框架在多样化的任务、领域和Transformer模型家族中的表现，包括使用BERT进行情感分析和长多类文档分类，使用XLM-R和AfroLM在低资源语言设置中进行仇恨言论检测，以及使用Masked Autoencoder Vision Transformer模型进行图像分类。在所有任务和架构中，CA-LIG提供了更忠实的属性，显示出对上下文依赖的更强敏感性，并产生了更清晰、更语义连贯的可视化结果，优于现有可解释性方法。这些结果表明，CA-LIG提供了更全面、上下文感知和可靠的Transformer决策解释，推动了深度神经网络的实用可解释性和概念理解。
英文摘要：
Transformer models achieve state-of-the-art performance across domains and tasks, yet their deeply layered representations make their predictions difficult to interpret. Existing explainability methods rely on final-layer attributions, capture either local token-level attributions or global attention patterns without unification, and lack context-awareness of inter-token dependencies and structural components. They also fail to capture how relevance evolves across layers and how structural components shape decision-making. To address these limitations, we proposed the \textbf{Context-Aware Layer-wise Integrated Gradients (CA-LIG) Framework}, a unified hierarchical attribution framework that computes layer-wise Integrated Gradients within each Transformer block and fuses these token-level attributions with class-specific attention gradients. This integration yields signed, context-sensitive attribution maps that capture supportive and opposing evidence while tracing the hierarchical flow of relevance through the Transformer layers. We evaluate the CA-LIG Framework across diverse tasks, domains, and transformer model families, including sentiment analysis and long and multi-class document classification with BERT, hate speech detection in a low-resource language setting with XLM-R and AfroLM, and image classification with Masked Autoencoder vision Transformer model. Across all tasks and architectures, CA-LIG provides more faithful attributions, shows stronger sensitivity to contextual dependencies, and produces clearer, more semantically coherent visualizations than established explainability methods. These results indicate that CA-LIG provides a more comprehensive, context-aware, and reliable explanation of Transformer decision-making, advancing both the practical interpretability and conceptual understanding of deep neural models.
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。
