Maintenance logs of industrial equipment record descriptive and unstructured operation and maintenance (O&M) information, which is the basis of reliability, availability, and maintainability investigations. However, the construction of failure knowledge graphs as a basis for understanding the failure and maintenance properties of systems is challenging due to the requirement of annotation efforts and domain knowledge. This article proposes a novel semi-supervised method for failure knowledge graph construction. Initially, a semantic module is proposed to extract hidden contextual information from maintenance records and identify corresponding failure modes. The semantic module is trained by unlabeled maintenance records with the assistance of the hard pseudo-label acquisition and the proposed self-training algorithm. Subsequently, a taxonomy induction module is presented to extract failure items and their relationships to construct failure knowledge graphs that provide decision support. The feasibility and superiority of the proposed method are validated by maintenance logs from real wind farms. Overall, the proposed method provides an effective tool for semantic information digitalization of well-cumulated industrial O&M data.

Product detection, which aims to localize products of interest in the advertising images, helps advance many potential E-commerce applications like product retrieval and recommendation. However, labeling a massive number of fine-grained product categories and accurate product boxes is costly and especially not practical since products are ever-changing on E-commerce websites. In this work, we step forward to train a fine-grained product detector solely supervised by the advertising captions, which are naturally available but often severely flawed and noisy. To reformulate the weakly supervised detection research into a real-world setting, we introduce a large-scale benchmark, namedCapProduct, where more than 80,000 product image-caption pairs are collected from E-commerce websites. The fine-grained nature of products and noisy captions inCapProductmake it intractable to excavate valid category labels to train a weakly supervised object detector. To tackle this challenge, we propose aCollaborativePseudo-LabelHarmonization (CoPLH) framework that harmonizes self-mined pseudo labels via modeling the global co-occurrence relationships of products. We construct a collaborative co-occurrence graph based on all training samples to improve the reliability of caption-predicted pseudo-labels as well as benefit the self-training procedure in a weakly supervised setting. Extensive experiments on theCapProductdataset demonstrate the effectiveness and the superiority of the proposed CoPLH over the state-of-the-art baselines.