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 2311.12418cs.CLcs.HC
收藏

生成式 Transformer 模型的可视化分析

Visual Analytics for Generative Transformer Models

University of British Columbia(不列颠哥伦比亚大学);University of Fraser Valley(弗雷泽河谷大学)

Raymond Li, Ruixin Yang, Wen Xiao, Ahmed AbuRaed, Gabriel Murray, Giuseppe Carenini

AI总结:

针对基于 transformer 的生成模型黑盒特性导致的可解释性挑战,提出一种新型可视化分析框架,支持编码器-解码器和仅解码器模型的生成与分类任务分析,并通过三个真实NLP案例验证了其可行性与实用性。

摘要:

虽然基于 transformer 的模型在各种分类和生成任务中取得了最先进的结果,但其黑盒特性使其在可解释性方面具有挑战性。在这项工作中,我们提出了一个新颖的可视化分析框架,以支持对基于 transformer 的生成网络的分析。与以往主要关注基于编码器的模型的工作不同,我们的框架是首批致力于支持对基于 transformer 的编码器-解码器模型和仅解码器模型进行生成和分类任务分析的框架之一。因此,我们提供了一个直观的概览,允许用户通过交互式可视化探索模型的不同层面。为了证明我们框架的可行性和实用性,我们基于现实世界的 NLP 研究问题展示了三个详细的案例研究。

英文摘要:

While transformer-based models have achieved state-of-the-art results in a variety of classification and generation tasks, their black-box nature makes them challenging for interpretability. In this work, we present a novel visual analytical framework to support the analysis of transformer-based generative networks. In contrast to previous work, which has mainly focused on encoder-based models, our framework is one of the first dedicated to supporting the analysis of transformer-based encoder-decoder models and decoder-only models for generative and classification tasks. Hence, we offer an intuitive overview that allows the user to explore different facets of the model through interactive visualization. To demonstrate the feasibility and usefulness of our framework, we present three detailed case studies based on real-world NLP research problems.

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。

↑