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 2311.14435cs.CVcs.AI
收藏

用于分析视觉 DNN 特征空间中概念分布的局部概念嵌入

Local Concept Embeddings for Analysis of Concept Distributions in Vision DNN Feature Spaces

Hochschule Anhalt(安哈尔特大学);Continental AG(Continental公司);University of Lübeck(吕贝克大学)

Georgii Mikriukov, Gesina Schwalbe, Korinna Bade

AI总结:

本文提出为每个样本生成 LoCE 的局部概念分析框架,用 GMM、聚类、检索和离群检测刻画 DNN 潜在空间中的概念分布,并在多数据集和多架构上验证其分析能力与分割性能。

摘要:

深入理解所学的潜在表征,对于在关键计算机视觉(CV)任务中验证深度神经网络(DNN)至关重要。因此,最先进的有监督基于概念的可解释人工智能(C-XAI)方法会将“汽车”等用户定义概念分别与 DNN 潜在空间中的单个向量(概念嵌入向量)关联起来。在概念分割情形下,这些向量可在属于某一概念的激活图像素与属于背景的激活图像素之间进行线性区分。然而,现有概念分割方法既无法捕捉隐式学习到的子概念(例如 DNN 可能将汽车拆分为“近处汽车”和“远处汽车”),也无法处理用户定义概念之间的重叠(例如“公交车”与“卡车”之间的重叠)。换言之,它们没有捕捉潜在空间中概念代表的完整分布。本研究首次表明,这些简化经常失效,而且分布信息对于理解 DNN 所学习的子概念、概念混淆和概念离群点等观念尤其有用。为支持探索所学的概念分布,我们提出一种新颖的局部概念分析框架。该框架并非在完整数据集上优化单个全局概念向量,而是为每个独立样本生成一个局部概念嵌入(LoCE)向量。我们利用 LoCE 构成的分布,通过拟合高斯混合模型(GMM)、层次聚类以及概念级信息检索和离群点检测来探索潜在概念分布。尽管该方法具有上下文敏感性,但其概念分割性能与全局基线相当。分析结果在三个数据集和六种不同的视觉 DNN 架构上获得,其中包括视觉 Transformer(ViT)。

英文摘要:

Insights into the learned latent representations are imperative for verifying deep neural networks (DNNs) in critical computer vision (CV) tasks. Therefore, state-of-the-art supervised Concept-based eXplainable Artificial Intelligence (C-XAI) methods associate user-defined concepts like ``car'' each with a single vector in the DNN latent space (concept embedding vector). In the case of concept segmentation, these linearly separate between activation map pixels belonging to a concept and those belonging to background. Existing methods for concept segmentation, however, fall short of capturing implicitly learned sub-concepts (e.g., the DNN might split car into ``proximate car'' and ``distant car''), and overlap of user-defined concepts (e.g., between ``bus'' and ``truck''). In other words, they do not capture the full distribution of concept representatives in latent space. For the first time, this work shows that these simplifications are frequently broken and that distribution information can be particularly useful for understanding DNN-learned notions of sub-concepts, concept confusion, and concept outliers. To allow exploration of learned concept distributions, we propose a novel local concept analysis framework. Instead of optimizing a single global concept vector on the complete dataset, it generates a local concept embedding (LoCE) vector for each individual sample. We use the distribution formed by LoCEs to explore the latent concept distribution by fitting Gaussian mixture models (GMMs), hierarchical clustering, and concept-level information retrieval and outlier detection. Despite its context sensitivity, our method's concept segmentation performance is competitive to global baselines. Analysis results are obtained on three datasets and six diverse vision DNN architectures, including vision transformers (ViTs).

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。

↑