This paper proposes a standard-free calibration transfer method for near-infrared (NIR) spectroscopy based on contrastive learning, aiming to address the challenge of calibration transfer caused by distribution shifts in spectral data across different instruments. We adapt and modify the original Contrastive Unpaired Translation (CUT) framework, initially designed for image domains, to accommodate one-dimensional spectral data, leveraging its strengths in one-way transfer and unsupervised learning. Experiments are conducted on a cross-device tobacco NIR dataset collected from two instruments. Quantitative results demonstrate that the proposed method achieves superior performance across multiple evaluation metrics, including Pearson correlation coefficient (PCC), cosine similarity (CS), mean absolute percentage error (MAPE) and root mean square error (RMSE). The proposed method outperforms existing standard-free methods such as Multiplicative Signal Correction (MSC) and Finite Impulse Response (FIR) filtering. Moreover, its performance, when compared to these methods, is closer to that of the existing standard calibration transfer method, Direct Standardization (DS). The method requires only standard-free data, significantly reducing the cost and complexity of calibration transfer in industrial applications where standard samples are scarce. These results highlight the potential of contrastive learning as a scalable and effective solution for real-world calibration transfer tasks.

This study proposes a standard-free calibration transfer (CT) method based on Cycle-Consistent Adversarial Networks (CycleGAN) to address the spectral differences between two near-infrared (NIR) spectrometers used for tobacco analysis under the same conditions. Developing standard-free CT methods is critical due to the limited availability of standard samples. Many traditional standard-free methods depend on restrictive assumptions, involve complex preprocessing procedures, or fail to fully leverage the available data, thereby limiting their applicability in practice. To overcome these limitations, this study proposes a CycleGAN-based transfer model. Specifically, the CycleGAN generator is adapted to one-dimensional architecture, and a Gated Convolutional Neural Network (Gated CNN) is integrated to enhance the ability of CycleGAN-based transfer model to capture long-range spectral dependencies. Moreover, an identity mapping loss is employed to constrain the transfer and minimize spectral trend distortions. Experimental results demonstrate that the proposed method not only addresses the limitations of traditional standard-free methods, but also achieves significantly lower spectral errors than both pre-transfer data and other standard-free methods such as Finite Impulse Response (FIR) filtering and Multiplicative Signal Correction (MSC). Although some structural consistency issues remain and require further optimization, CycleGAN-based transfer model represents a promising approach for cross-device NIR tobacco data transfer, demonstrating technical benefits and strong potential for deployment in complex industrial environments.

As a key component of industrial equipment, bearings are prone to failure under complex operating conditions. Bearing fault diagnosis can detect potential dangers at an early stage, thus ensuring the stability and efficiency of production. Existing research mainly focuses on the structural improvement of the fault classification model and often neglects the optimization of the algorithm parameters. To address this deficiency, this paper proposed a hybrid convolutional neural network-long short-term memory (CNN-LSTM) model for bearing fault diagnosis. The model parameters were optimized using the improved whale optimization algorithm (IWOA). Experimental results show that the proposed method has superior performance in bearing fault diagnosis and has broad application prospects.

Knowledge Graph Completion (KGC) aims to enrich and complete the knowledge graph by discovering missing information from existing fact triples. However, existing KGC methods often overlook the utilization of structured knowledge within the knowledge base. In this paper, we propose a novel Large Language Models-based Knowledge Graph Completion framework, called SIKGC, which builds the structural information prompt to assist the knowledge graph completion tasks. Specifically, we arrange the triples in the knowledge graph as the sequences of text. By fusing the descriptions of entities, relations and their structural information as task-aware prompts, we input such prompts into large language models and regard the responses as prediction tasks. The experimental results on various public datasets show that the proposed method outperforms all baseline methods for the three knowledge completion tasks and attains state-of-the-art in triple classification. We also demonstrate that fine-tuning the smaller large language models (e.g., Baichuan2-13B, LLaMA2-13B, ChatGLM3-6B) with relevant data markedly enhances their KGC capabilities and significantly outperforms GPT-4.

Abstract Defect detection in cigarette packaging is a crucial process for ensuring product quality meets industry standards within the tobacco manufacturing sector. Defect detection methods based on deep learning have significantly enhanced efficiency. However, challenges remain in addressing issues such as blurred boundary textures of cigarette pack defects and the complex differentiability and similarity among defects. Consequently, this study presents the design of a defect detector for cigarette packs that is sensitive to detail features, named AMC‐YOLO. Initially, an Adaptive Spatial Weight Perception (ASWP) module is designed to emphasize local information from different regions during the downsampling process and integrate effective features. Additionally, a Multidimensional Aggregation Radiative Feature Pyramid Networks (MARFPN) is proposed to aggregate multi‐scale semantic information across dimensions and relay it back to various levels within the network to facilitate the learning of more refined feature. Lastly, a Cross‐Layer Collaborative Detection Head (CLCDH) is introduced to further weight and fuse the contextual information between local and global aspects. Experimental results demonstrate that AMC‐YOLO outperforms state‐of‐the‐art methods on the Cigarette Pack Defect and GC10‐DET datasets, exhibiting the highest detection precision and excellent generalization. These findings highlight the significant potential for the application of AMC‐YOLO in cigarette pack defect detection.