Purpose: This study proposed a novel Local Reference Semantic Code (LRSC) network for automatic breast ultrasound image classification with few labeled data. Methods: In the proposed network, the local structure extractor is firstly developed to learn the local reference which describes common local characteristics of tumors. After that, a two-stage hierarchical encoder is developed to encode the local structures of lesion into the high-level semantic code. Based on the learned semantic code, the self-matching layer is proposed for the final classification. Results: In the experiment, the proposed method outperformed traditional classification methods and AUC (Area Under Curve), ACC (Accuracy), Sen (Sensitivity), Spec (Specificity), PPV (Positive Predictive Values), and NPV(Negative Predictive Values) are 0.9540, 0.9776, 0.9629, 0.93, 0.9774 and 0.9090, respectively. In addition, the proposed method also improved matching speed. Conclusions: LRSC-network is proposed for breast ultrasound images classification with few labeled data. In the proposed network, a two-stage hierarchical encoder is introduced to learn high-level semantic code. The learned code contains more effective high-level classification information and is simpler, leading to better generalization ability.

With the rapid advancement of industrial intelligence, equipment predictive maintenance (PdM) has emerged as a critical enabler for intelligent operation and maintenance (O&M) systems. However, conventional machine learning-based PdM methods are inherently constrained by scarce labeled fault samples, weak cross-condition generalization, and a lack of advanced causal reasoning capabilities in complex industrial process environments. Recent advances in large language models (LLMs) and generative artificial intelligence (GAI) have created unprecedented opportunities to fundamentally overcome these persistent PdM limitations. This paper presents a comprehensive review and future perspective on LLM applications for PdM in industrial process systems and proposes a hierarchical “Capability-Domain-Method (CDM)” taxonomy to establish a unified analytical framework for this fast-growing research field. Based on a systematic analysis of the 103 core research papers published between 2016 and 2026, this work comprehensively summarizes the latest research on LLMs for predictive maintenance from three core technical dimensions: data augmentation, model adaptation, and knowledge reasoning and decision-making. Data augmentation methods generate high-fidelity fault samples and degradation patterns to mitigate industrial data scarcity and class imbalance. Model adaptation techniques empower general-purpose pre-trained LLMs to adapt to specific equipment and complex industrial environments. Knowledge reasoning and decision-making approaches integrate external industrial domain knowledge, autonomous reasoning mechanisms, and maintenance decision processes to improve the performance of fault diagnosis, root cause analysis, and maintenance decision-making. Furthermore, this paper discusses key challenges related to operational reliability, deep domain adaptation, unified evaluation standards, and unified deployment efficiency. Finally, future research directions are advised, including physics-informed large models, industrial domain-specific models, autonomous maintenance agents, and standardized open benchmarking platforms, to accelerate the paradigm shift in PdM from data-driven pattern to knowledge-driven and autonomous decision-making.