Loss Function Research Articles

Hypercube Pooling for Visual Semantic Embedding

Visual Semantic Embedding ( VSE ) is a primary model for cross-modal retrieval, wherein the global feature aggregator is a crucial component of the VSE model. In recent research, the General Pooling Operator ( GPO ) aggregator, which weighs the features reconstructed from the local feature set to aggregate, facilitates the related models to achieve good retrieval performance. However, the reason for the effectiveness remains to be explored. To enhance the rationality of aggregator designs, we analyze the reason from the perspective of feature space. Indeed, for each data, the local feature set forms a hypercube containing abundant data information, and the feature learned by GPO measures the hypercube, thereby representing the data. The geometric structure of the hypercube implies that the set containing all points within the hypercube is a convex set, so the feature learned by weighted aggregation is an interior point of the hypercube. However, using the interior point to measure the hypercube leads to some problems in feature representation and model optimization, as well as the reduction of retrieval efficiency caused by weight computation. For example, the related pair’s features may be far, while the unrelated ones may be close. To measure the hypercube more clearly and alleviate the problems mentioned above, we propose Hypercube Pooling ( HCP ) aggregator. Specifically, HCP concatenates the Max and Min Pooling features as the global features. This aggregation method has multiple advantages, e.g., the learned global feature represents all hyperplanes of the hypercube that contain critical information and hypercube geometric structure. Moreover, HCP adds normalization-before-concatenation and reduces the usual setting of margin in the loss function by half to avoid gradient loss caused by the difference in the feature value and dimensionality doubling. The experimental results on the Flickr30K and MSCOCO datasets show that the HCP model has excellent performance with high efficiency, confirming the correctness of the spatial analysis and the effectiveness of the HCP aggregator.

A comparison of artificial intelligence-enhanced electrocardiography approaches for prediction of time-to-mortality using electrocardiogram images

Abstract Background Most artificial intelligence-enhanced ECG (AI-ECG) models used to predict adverse events including death require that the ECGs be stored digitally. However, the majority of clinical facilities worldwide store ECGs as images. Methods 1,163,401 ECGs (189,539 patients) from a secondary care dataset were available as both natively digital traces and PDF images. A digitisation pipeline extracted signals from PDFs. Separate 1D convolutional neural network (CNN) models were trained on natively digital or digitised ECGs, with a discrete-time survival loss function to predict time-to-mortality. A 2D CNN model was trained on 310x868 pixel ECG images. External validation was performed in 958,954 ECGs (645,373 patients) from a Brazilian primary care cohort and 1022 ECGs (1022 patients) from a Chagas disease cohort. Results The image 2D CNN model and digitised 1D CNN model performed comparably to natively digital 1D CNN model in internal (C-index 0.780 (0.779-0.781), 0.772 (0.771-0.774) and 0.775 (0.774-0.776 respectively) and external validation. Models trained on natively digital 1D ECGs had comparable performance when applied to digitised 1D ECGs (C-index 0.773 (0.771-0.774). Conclusion Both the image 2D CNN and digitised 1D CNN enable mortality prediction from ECG images, with comparable performance to natively digital 1D CNN. Models trained on natively digital 1D ECGs can also be applied to digitised 1D ECGs, without any significant loss in performance. This works allows AI-ECG mortality prediction to be applied in diverse global settings lacking digital ECG infrastructure.

Fabric defect detection method based on multi-scale fusion attention mechanisms

Abstract Fabric defect detection is extremely important for the development of the textile industry, but the existing traditional image processing algorithms are not good enough to detect fabric defects, and the detection efficiency and accuracy of the classical deep learning model is not satisfactory, so this paper proposes an improved fabric defect detection method based on multi-scale fusion of attention mechanism YOLOv7-PCBS. Based on the YOLOv7 network structure, some of the standard convolutions of the backbone network are replaced with Partial Convolution modules, which reduces the amount of network computation and improves the network detection speed; add coordinate attention to enhance the ability of extracting the positional features of tiny defects in fabrics; reconfiguration of the SPPCSPC module to improve small target detection; optimization of Bidirectional Feature Pyramid Network (BiFPN) and design of Tiny- BiFPN for simple and fast multi-scale feature fusion; finally, a novel loss function SIoU with angular loss is introduced to facilitate the fitting of the true and predicted frames and enhance the accuracy of defect prediction. The results show that the algorithm achieves a mAP value of 94.4% on the detection of defects in solid-colored fabrics of six denim materials, which is an improvement of 15.1% compared to the original YOLOv7 algorithm, while the model achieves a frame rate of 59.5 per second. Compared with other traditional deep learning algorithms SSD and Faster-RCNN, the detection accuracies are improved by 21.6% and 15.2%, and the FPS values are improved by 78.1% and 101.0%, respectively. Therefore, the YOLOv7-PCBS fabric defect detection algorithm proposed in this paper makes the fabric defect detection results more accurate while realizing lightweight, which provides an important technical reference for the subsequent improvement of textile quality.

Research on coal gangue recognition method based on SFD-YOLOv5s

ABSTRACT Aiming at the current coal gangue recognition model’s difficulty in recognizing aggregated coal gangue targets, which leads to the problems of misdetections, omissions, and inaccurate localization, we propose a coal gangue recognition method based on the SFD-YOLOv5s model. The SE attention mechanism is improved in incorporating maximum pooling and introduced into the backbone network of the YOLOv5s model to improve the model’s ability to extract the critical feature information of the coal gangue. Furthermore, to improve the recognition ability of the YOLOv5s model for aggregated targets, we introduce the Focal-EIoU loss function and the Dyhead. The experimental results show that the overall performance of the SFD-YOLOv5s model has surpassed other mainstream object detection models of the YOLO series. The SFD-YOLOv5s model has the highest precision, recall, and mean average precision. The precision reaches 97.9%, the recall is 96.2%, and the mean average precision is 94.5%, which is 8.7%, 8.9%, and 9.0% higher compared with the YOLOv5s model, respectively. The FPS is 96.15, which can fully meet the demand for real-time detection of coal gangue. Meanwhile, the SFD-YOLOv5s model has the best recognition effect of aggregated coal gangue targets.

Mechanical, electronic, optical, and thermoelectric properties of Rb2TlGaX6 (X = F, I) lead free double perovskites: A first-principles study

Abstract Halide perovskites play an important role in the renewable energy industry and are the top substitutes for Pb-halide perovskites. This study investigates the structural, mechanical, electronic, optical, and thermoelectric properties of the double halide perovskites Rb2TlGaX6 (X = F, I) using density functional theory (DFT) at WIEN2k interfaces. The structural, mechanical, and thermal stabilities are evaluated using the tolerance factor, Born-Huang criteria, and negative formation energy. The obtained lattice constants are 9.31 Å and 24.93 Å, for Rb2TlGaF6 and Rb2TlGaI6 respectively. Rb2TlGaF6 is brittle and Rb2TlGaI6 is ductile. Both compounds exhibit anisotropic behavior. The indirect semiconducting nature (L-X) is demonstrated by the estimated bandgap values of 6.05 eV (4.31 eV) and 1.06 eV (0.52 eV), for Rb2TlGaF6 and Rb2TlGaI6 respectively, using the TB-mBJ (PBE-GGA) potential. Distinctive optical characteristics are studied, including the dielectric function, loss function, conductivity, reflectivity, refractive index, and absorption coefficient. Based on the investigation, both materials exhibit excellent optical conductivity, high light absorption throughout the UV–visible spectrum, and low reflection (within 20% for Rb2TlGaF6 and 40% for Rb2TlGaI6). Lastly, various thermoelectric properties are explored by varying the temperature. Both compounds show a high Seebeck coefficient and low thermal conductivity. The maximum figure of merit (ZT) obtains of 0.61 and 0.81, for Rb2TlGaF6 and Rb2TlGaI6, respectively. These results demonstrate the suitability of Rb2TlGaF6 and Rb2TlGaI6 for thermoelectric and photovoltaic applications.

A Helium Speech Correction Method Based on Generative Adversarial Networks

The distortion of helium speech caused by helium−oxygen gas mixtures significantly impacts the safety and communication efficiency of saturation divers. Although existing correction methods have shown some effectiveness in improving the intelligibility of helium speech, challenges remain in enhancing clarity and high−pitch correction. To address the issue of degraded speech quality post−correction, a novel helium speech correction method based on generative adversarial networks (GANs) is proposed. Firstly, a new helium speech dataset is introduced, which includes isolated words and continuous speech in both Chinese and English. By training and testing on both isolated words and continuous passages, the correction capability of the model can be accurately evaluated. Secondly, a new evaluation system for helium speech correction is proposed, which partially fills the gap in current helium speech evaluation metrics. This system uses comprehensive similarity to evaluate the similarity of keywords at the sentence level, thus assessing the correction results of helium speech from both word and sentence dimensions. Lastly, a GAN−based helium speech correction method is designed. This method solves the problems of pitch period distortion and formant shift in helium speech by introducing an adaptive speech segmentation algorithm and a fusion loss function and significantly improves the clarity and intelligibility of corrected helium speech. The experimental results show that the corrected helium speech is improved in clarity and intelligibility, which shows its practical value and application potential.

Mask-guided network for finger vein feature extraction and biometric identification

The problems of complex background, low quality of finger vein images, and poor discriminative features have been the bottleneck of feature extraction and finger vein recognition. To this end, we propose a feature extraction algorithm based on the open-set testing protocol. In order to eliminate the interference of irrelevant areas, this paper proposes the idea of segmentation-assisted classification, that is, using the rough mask of the finger vein to constrain the feature learning process so that the network can focus on the vein area and learn greater weight for the vein. Specifically, the feature maps of the shallow layers of the network are first sent to the feature pyramid module to fuse the primary features of different scales, which are then sent to the spatial attention module to obtain the spatial weight map of the image. Based on the results of several classical vein skeleton extraction algorithms, a weighting method is used to obtain a more accurate mask to constrain the learning of the spatial weight map. Finally, a hybrid loss function combining triplet loss and cross-entropy loss is used to reduce the distance between feature vectors of the same categories and increase the distance between feature vectors of different categories in the Euclidean space, thereby improving feature discriminability. Good recognition results were achieved on the three public data sets of SDUMLA, MMCBNU, and FVUSM, and the values of equal error rate (EER) on them are as low as 2.50%, 0.20%, and 0.14%, respectively.

A highly efficient tunnel lining crack detection model based on Mini-Unet.

The accurate detection of tunnel lining cracks and prompt identification of their primary causes are critical for maintaining tunnel availability. The advancement of deep learning, particularly in the domain of convolutional neural network (CNN) for image segmentation, has made tunnel lining crack detection more feasible. However, the CNN-based technique for tunnel lining crack detection commonly prioritizes increasing algorithmic complexity to enhance detection accuracy, posing a challenge in balancing the accuracy of detection and the efficiency of the algorithm. Motivated by the superior performance of Unet in image segmentation, this paper proposes a lightweight tunnel lining crack detection model named Mini-Unet, which refined the Unet architecture and utilized depthwise separable convolutions (DSConv) to replace some standard convolution layers. In the optimization of the proposed model parameters, applying a hybrid loss function that integrated dice loss and cross-entropy loss effectively tackled the imbalance between crack and background categories. Several models were set up to contrast with Mini-Unet and the experimental results were analyzed. Mini-Unet achieves a mean intersection over union (MIoU) of 60.76%, a mean precision of 84.18%, and a frame per second (FPS) of 5.635, respectively. Mini-Unet outperforms several mainstream models, enabling rapid detection while maintaining identified accuracy and facilitating the practical application of AI power for real-time tunnel lining crack detection.

Stochastic Gradient Descent with Pre-Conditioned Polyak Step-Size

Stochastic Gradient Descent (SGD) is one of the many iterative optimization methods that are widely used in solving machine learning problems. These methods display valuable properties and attract researchers and industrial machine learning engineers with their simplicity. However, one of the weaknesses of this type of methods is the necessity to tune learning rate (step-size) for every loss function and dataset combination to solve an optimization problem and get an efficient performance in a given time budget. Stochastic Gradient Descent with Polyak Step-size (SPS) is a method that offers an update rule that alleviates the need of fine-tuning the learning rate of an optimizer. In this paper, we propose an extension of SPS that employs preconditioning techniques, such as Hutchinson’s method, Adam, and AdaGrad, to improve its performance on badly scaled and/or ill-conditioned datasets.

MaskDGNets: Masked-attention guided dynamic graph aggregation network for event extraction.

Considering that the traditional deep learning event extraction method ignores the correlation between word features and sequence information, it cannot fully explore the hidden associations between events and events and between events and primary attributes. To solve these problems, we developed a new framework for event extraction called the masked attention-guided dynamic graph aggregation network. On the one hand, to obtain effective word representation and sequence representation, an interaction and complementary relationship are established between word vectors and character vectors. At the same time, a squeeze layer is introduced in the bidirectional independent recurrent unit to model the sentence sequence from both positive and negative directions while retaining the local spatial details to the maximum extent and establishing practical long-term dependencies and rich global context representations. On the other hand, the designed masked attention mechanism can effectively balance the word vector features and sequence semantics and refine these features. The designed dynamic graph aggregation module establishes effective connections between events and events, and between events and essential attributes, strengthens the interactivity and association between them, and realizes feature transfer and aggregation on graph nodes in the neighborhood through dynamic strategies to improve the performance of event extraction. We designed a reconstructed weighted loss function to supervise and adjust each module individually to ensure the optimal feature representation. Finally, the proposed MaskDGNets framework is evaluated on two baseline datasets, DuEE and CCKS2020. It demonstrates its robustness and event extraction performance, with F1 of 81.443% and 87.382%, respectively.

Fault diagnosis method for elevator carriages based on temporal generative federated distillation

Addressing the ‘data silo’ issue among different elevator operating units and the temporal correlations in elevator vibration signals, a novel small-sample fault diagnosis method for elevator carriages based on temporal generative federated distillation is proposed. This method incorporates a temporal generative adversarial network into Federated Distillation via Generative Learning (FedGen). FedGen combines federated learning, knowledge distillation, and generative models to enhance model aggregation efficiency while mitigating data heterogeneity. However, the original generative model struggles to maintain dynamic correlations between signals when extracting temporal features. Therefore, an improved Time Series Generative Adversarial Networks (TimeGAN) model is introduced, substituting the initial logarithmic loss function with a least squares error function, thereby enhancing training stability and data quality. This approach eliminates the need for proxy datasets in knowledge distillation, avoiding the loss of temporal information during central server feature extraction. Simulation results demonstrate that this method enables data sharing while protecting data privacy, and enhances model generalization capabilities.

Lightweight enhanced YOLOv8n underwater object detection network for low light environments.

In response to the challenges of target misidentification, missed detection, and other issues arising from severe light attenuation, low visibility, and complex environments in current underwater target detection, we propose a lightweight low-light underwater target detection network, named PDSC-YOLOv8n. Firstly, we enhance the input images using the improved Pro MSRCR algorithm for data augmentation. Secondly, we replace the traditional convolutions in the backbone and neck networks of YOLOv8n with Ghost and GSConv modules respectively to achieve lightweight network modeling. Additionally, we integrate the improved DCNv3 module into the C2f module of the backbone network to enhance the capability of target feature extraction. Furthermore, we introduce the GAM into the SPPF and incorporate the CBAM attention mechanism into the last layer of the backbone network to enhance the model's perceptual and generalization capabilities. Finally, we optimize the training process of the model using WIoUv3 as the loss function. The model is successfully deployed on an embedded platform, achieving real-time detection of underwater targets on the embedded platform. We first conduct experiments on the RUOD underwater dataset. Meanwhile, we also utilized the Pascal VOC2012 dataset to evaluate our approach. The mAP@0.5 and mAP@0.5:0.95 of the original YOLOv8n algorithm on RUOD dataset were 79.6% and 58.2%, respectively, and the PDSC -YOLOv8n algorithm mAP@0.5 and mAP@0.5:0.95 can reach 86.1% and 60.8%. The number of parameters of the model is reduced by about 15.5%, the detection accuracy is improved by 6.5%. The original YOLOv8n algorithm was 73% and 53.2% mAP@0.5 and mAP@0.5:0.95 on the Pascal VOC dataset, respectively. The PDSC-YOLOv8n algorithm mAP@0.5 and mAP@0.5:0.95 were 78.5% and 57%, respectively. The superior performance of PDSC-YOLOv8n indicates its effectiveness in the field of underwater target detection.

A novel deep learning‐based technique for efficient characterization of engineered cementitious composites cracks for durability assessment

AbstractEngineered Cementitious Composites also known as Strain‐hardening cementitious composites (SHCCs) has unique cracking patterns like cracks that have tiny widths and showcase high density. All of this makes it difficult and laborious to compute crack parameters from crack patterns. Unfortunately, this is an essential part of assessing durability and micromechanical modeling. SHSnet is developed to perform end‐to‐end semantic segmentation of SHCC cracks. SHSnet is efficient, attention based deep encoder‐decoder network with large receptive field. Loss function based on Tversky function were used for training the model. SHSnet with loss function shows promising result with mPrecision, mF1Score and mIoU of 0.87, 0.84 and 0.83 respectively for complex SHCC cracks while requiring at least an order of fewer computational parameters than those in the literature. An image processing unit is then used to estimate the width, number, and length of the cracks from the segmentation mask. Test results show that the computed crack parameters with SHSnet are exactly the same as that computed with an optical microscope but require ~100× less time. Results demonstrate that SHSnet works equally well in SHCCs with different surface textures, crack density, and widths; the final result was far superior to a conventional technique. This technique also shows promising results in an automatic evaluation of crack parameters relevant to durability and visualizing crack patterns even in the presence of artifacts during progressive testing. The results also demonstrate the necessity to accurately and densely calculate crack length and maximum crack width; else the durability results are expected to be significantly more conservative than the actual value.

Single Well Production Prediction Model of Gas Reservoir Based on CNN-BILSTM-AM

In the prediction of single-well production in gas reservoirs, the traditional empirical formula of gas reservoirs generally shows poor accuracy. In the process of machine learning training and prediction, the problems of small data volume and dirty data are often encountered. In order to overcome the above problems, a single-well production prediction model of gas reservoirs based on CNN-BILSTM-AM is proposed. The model is built by long-term and short-term memory neural networks, convolutional neural networks and attention modules. The input of the model includes the production of the previous period and its influencing factors. At the same time, the fitting production and error value of the traditional gas reservoir empirical formula are introduced to predict the future production data. The loss function is used to evaluate the deviation between the predicted data and the real data, and the Bayesian hyperparameter optimization algorithm is used to optimize the model structure and comprehensively improve the generalization ability of the model. Three single wells in the Daniudi D28 well area were selected as the database, and the CNN-BILSTM-AM model was used to predict the single-well production. The results show that compared with the prediction results of the convolutional neural network (CNN) model, long short-term memory neural network (LSTM) model and bidirectional long short-term memory neural network (BILSTM) model, the error of the CNN-BILSTM-AM model on the test set of three experimental wells is reduced by 6.2425%, 4.9522% and 3.0750% on average. It shows that on the basis of coupling the empirical formula of traditional gas reservoirs, the CNN-BILSTM-AM model meets the high-precision requirements for the single-well production prediction of gas reservoirs, which is of great significance to guide the efficient development of oil fields and ensure the safety of China’s energy strategy.

Abstract B053: NucAE: Autoencoder-based enhancement of nucleosome occupancy signals from low-coverage cfDNA sequencing

Abstract Introduction: Cell-free DNA (cfDNA) is emerging as a valuable cancer biomarker, enabling both the detection and epigenetic characterization of malignancies through sequencing. cfDNA sequencing can reveal cancer-specific nucleosome occupancy patterns, which provide insights into tumor type and differentiation status, which define prognosis and therapy recommendations. Low-coverage whole-genome cfDNA sequencing is a cost-efficient approach to assay multiple cfDNA samples, however, the low coverage limits its sensitivity. We developed NucAE, a denoising autoencoder model, to enhance nucleosome occupancy signals from low-coverage cfDNA-sequencing. Methods: We designed NucAE to estimate nucleosome occupancy genome- wide from low-coverage sequencing data. To train the model, we undersampled high-coverage cfDNA-sequencing data and produced high- and low-coverage sample pairs at varying coverages (1x-90x). Inputs to NucAE include cfDNA fragment data and the corresponding genomic nucleotide sequence, with the model leveraging a symmetrical architecture comprising two convolutional layers in both the encoder and decoder, connected by ReLU activations. Mean squared error relative to the high-coverage signal was employed as the loss function. To evaluate the model independently, we performed peak detection from the nucleosome occupancy signals. We conducted a grid search to optimize hyperparameters. Results: We found that NucAE enhances the signal to noise ratio by detecting periodicities in nucleosome occupancy data. Denoising was improved by supplementing the cfDNA fragment signals with the genomic nucleotide sequences as input, which demonstrates that the model learned the nucleotide- sequence dependence of nucleosome positioning. Independent validation through peak detection on previously unseen samples demonstrated a significant improvement (p<0.001) in accuracy compared to raw signals across all tested coverages (1x-45x). Remarkably, at extremely low coverages (1x-2x), denoised signals allowed for more accurate peak detection than raw signals from 10-20x deeper sequencing. Conclusions: By using advanced machine-learning, we were able to enhance low-coverage cfDNA sequencing data to assay nucleosome positions at a nucleotide resolution. NucAE achieves an order of magnitude improvement over the low- coverage signals and greatly improves the utility of cfDNA-sequencing data to identify cancer- specific nucleosome occupancy, while preserving its low cost. Our model is the first to use autoencoders for genome-wide signal enhancement with potential use cases in many other areas of genomics. Citation Format: Zsolt Balázs, Jean Radig, Noah Wolford, Manuel Schürch, Michael Krauthammer. NucAE: Autoencoder-based enhancement of nucleosome occupancy signals from low-coverage cfDNA sequencing [abstract]. In: Proceedings of the AACR Special Conference: Liquid Biopsy: From Discovery to Clinical Implementation; 2024 Nov 13-16; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2024;30(21_Suppl):Abstract nr B053.

Building Change Detection Network Based on Multilevel Geometric Representation Optimization Using Frame Fields

To address the challenges of accurately segmenting irregular building boundaries in complex urban environments faced by existing remote sensing change detection methods, this paper proposes a building change detection network based on multilevel geometric representation optimization using frame fields called BuildingCDNet. The proposed method employs a multi-scale feature aggregation encoder–decoder architecture, leveraging contextual information to capture the characteristics of buildings of varying sizes in the imagery. Cross-attention mechanisms are incorporated to enhance the feature correlations between the change pairs. Additionally, the frame field is introduced into the network to model the complex geometric structure of the building target. By learning the local orientation information of the building structure, the frame field can effectively capture the geometric features of complex building features. During the training process, a multi-task learning strategy is used to align the predicted frame field with the real building outline, while learning the overall segmentation, edge outline, and corner point features of the building. This improves the accuracy of the building polygon representation. Furthermore, a discriminative loss function is constructed through multi-task learning to optimize the polygonal structured information of the building targets. The proposed method achieves state-of-the-art results on two commonly used datasets.

WKNN-Based Wi-Fi Fingerprinting with Deep Distance Metric Learning via Siamese Triplet Network for Indoor Positioning

Weighted k-nearest neighbor (WKNN)-based Wi-Fi fingerprinting is popular in indoor location-based services due to its ease of implementation and low computational cost. KNN-based methods rely on distance metrics to select the nearest neighbors. However, traditional metrics often fail to capture the complexity of indoor environments and have limitations in identifying non-linear relationships. To address these issues, we propose a novel WKNN-based Wi-Fi fingerprinting method that incorporates distance metric learning. In the offline phase, our method utilizes a Siamese network with a triplet loss function to learn a meaningful distance metric from training fingerprints (FPs). This process employs a unique triplet mining strategy to handle the inherent noise in FPs. Subsequently, in the online phase, the learned metric is used to calculate the embedding distance, followed by a signal-space distance filtering step to optimally select neighbors and estimate the user’s location. The filtering step mitigates issues from an overfitted distance metric influenced by hard triplets, which could lead to incorrect neighbor selection. We evaluate the proposed method on three benchmark datasets, UJIIndoorLoc, Tampere, and UTSIndoorLoc, and compare it with four WKNN models. The results show a mean positioning error reduction of 3.55% on UJIIndoorLoc, 16.21% on Tampere, and 16.49% on UTSIndoorLoc, demonstrating enhanced positioning accuracy.

Trainability barriers and opportunities in quantum generative modeling

AbstractQuantum generative models provide inherently efficient sampling strategies and thus show promise for achieving an advantage using quantum hardware. In this work, we investigate the barriers to the trainability of quantum generative models posed by barren plateaus and exponential loss concentration. We explore the interplay between explicit and implicit models and losses, and show that using quantum generative models with explicit losses such as the KL divergence leads to a new flavor of barren plateaus. In contrast, the implicit Maximum Mean Discrepancy loss can be viewed as the expectation value of an observable that is either low-bodied and provably trainable, or global and untrainable depending on the choice of kernel. In parallel, we find that solely low-bodied implicit losses cannot in general distinguish high-order correlations in the target data, while some quantum loss estimation strategies can. We validate our findings by comparing different loss functions for modeling data from High-Energy-Physics.

MDD-DETR: Lightweight Detection Algorithm for Printed Circuit Board Minor Defects

PCBs (printed circuit boards) are the core components of modern electronic devices, and inspecting them for defects will have a direct impact on the performance, reliability and cost of the product. However, the performance of current detection algorithms in identifying minor PCB defects (e.g., mouse bite and spur) still requires improvement. This paper presents the MDD-DETR algorithm for detecting minor defects in PCBs. The backbone network, MDDNet, is used to efficiently extract features while significantly reducing the number of parameters. Simultaneously, the HiLo attention mechanism captures both high- and low-frequency features, transmitting a broader range of gradient information to the neck. Additionally, the proposed SOEP neck network effectively fuses scale features, particularly those rich in small targets, while INM-IoU loss function optimization enables more effective distinction between defects and background, further improving detection accuracy. Experimental results on the PCB_DATASET show that MDD-DETR achieves a 99.3% mAP, outperforming RT-DETR by 2.0% and reducing parameters by 32.3%, thus effectively addressing the challenges of detecting minor PCB defects.

LCANet: a model for analysis of students real-time sentiment by integrating attention mechanism and joint loss function

LCANet: a model for analysis of students real-time sentiment by integrating attention mechanism and joint loss function