Performance In Tasks Research Articles

Artificial Intelligence and Convolutional Neural Networks-Driven Detection of Micro and Macro Metastasis of Cutaneous Melanoma to the Lymph Nodes.

Lymph node (LN) assessment is a critical component in the staging and management of cutaneous melanoma. Traditional histopathological evaluation, supported by immunohistochemical staining, is the gold standard for detecting LN metastases. However, the process is labor-intensive, requiring the analysis of multiple tissue levels, which increases both time and cost. With the growing integration of artificial intelligence (AI) into clinical workflows, there is potential to streamline this process, enhancing efficiency and accuracy. This study analyzed 53 WSIs (12 negative samples). Images are all derived from hematoxylin-eosin (H&E) stains and were uploaded to the Orca AI cloud-based platform (SpatialX Diagnostics, Inc.), a commercially available platform specifically designed for AI pathology. An AI-driven model was developed and trained to detect melanoma metastases directly from H&E histopathological images, bypassing additional immunohistochemical staining. The algorithm was designed to identify the presence of metastases and classify tumor deposits based on size, specifically distinguishing deposits greater than 0.1 mm and less than 1.0 mm in diameter, which are critical thresholds for prognostic evaluation. The model was validated on 9 WSIs (2 negative) that were not seen by the model. For every WSI in the validation set, one of the following 4 classes was assigned: normal, metastasis (<0.1 mm), metastasis (0.1-1 mm), and metastasis (>1 mm). The class corresponding to the highest metastasis size range was assigned to the whole sample. The results were then examined by a board-certified pathologist. The AI algorithm demonstrated high accuracy in detecting LN metastases in patients with melanoma. It effectively identified and classified tumor deposits with a specificity of 0.91 and a sensitivity of 0.74. The model also distinguishes between smaller deposits (>0.1 mm) and larger deposits (>1.0 mm). This stratification is essential for accurate staging, prognosis determination, and treatment planning, highlighting the algorithm's potential clinical value. When evaluating the model performance in the WSI classification tasks, the model showed high agreement with the pathologist's classification, correctly identifying 7 WSIs as metastasis (>1 mm) and labeling 1 normal WSI as metastasis (0.7 mm). The study's findings underscore the potential of AI in revolutionizing the detection of melanoma metastasis in lymph nodes. By significantly reducing the reliance on time-consuming and costly immunohistochemical staining, AI-driven tools can streamline diagnostic workflows, improve accuracy, and potentially enhance patient outcomes. As AI technology continues to evolve, its application in melanoma management could become a cornerstone of modern pathology, offering a powerful adjunct to traditional diagnostic methods.

Beyond Binary Decisions: Evaluating the Effects of AI Error Type on Trust and Performance in AI-Assisted Tasks.

ObjectiveWe investigated how various error patterns from an AI aid in the nonbinary decision scenario influence human operators' trust in the AI system and their task performance.BackgroundExisting research on trust in automation/autonomy predominantly uses the signal detection theory (SDT) to model autonomy performance. The SDT classifies the world into binary states and hence oversimplifies the interaction observed in real-world scenarios. Allowing multi-class classification of the world reveals intriguing error patterns previously unexplored in prior literature.MethodThirty-five participants completed 60 trials of a simulated mental rotation task assisted by an AI with 70-80% reliability. Participants' trust in and dependence on the AI system and their performance were measured. By combining participants' initial performance and the AI aid's performance, five distinct patterns emerged. Mixed-effects models were built to examine the effects of different patterns on trust adjustment, performance, and reaction time.ResultsVarying error patterns from AI impacted performance, reaction times, and trust. Some AI errors provided false reassurance, misleading operators into believing their incorrect decisions were correct, worsening performance and trust. Paradoxically, some AI errors prompted safety checks and verifications, which, despite causing a moderate decrease in trust, ultimately enhanced overall performance.ConclusionThe findings demonstrate that the types of errors made by an AI system significantly affect human trust and performance, emphasizing the need to model the complicated human-AI interaction in real life.ApplicationThese insights can guide the development of AI systems that classify the state of the world into multiple classes, enabling the operators to make more informed and accurate decisions based on feedback.

Effects of trust and perceived self-efficacy on aided judgement: an experiment on spelling and grammar checking

ABSTRACT We report an experimental study in which participants make a series of linguistic judgements with help from a fictional text-editing aid. The study illuminates how acceptance of a software aid's advice might be affected by, and in turn affects, users’ trust and confidence. We measure the propensity to accept automated advice by adapting Signal Detection Theory (SDT), and particularly the construct of bias. Our study demonstrates a significant difference in bias between groups who interact with aids that perform at different levels of reliability. Participants’ pre-task trust in similar systems also affected bias towards the aid, but we found no evidence for such an effect of prior perceived self-efficacy. Users show an awareness of their own performance, demonstrated by higher self-reported confidence in correct responses than in incorrect ones. Further, participants using a more reliable aid are more confident when accepting than when rejecting the aid's advice. Overall, participants were overconfident in their judgements. Our research demonstrates opportunities of using an experimental paradigm, analysed with SDT, to study aspects of performance, trust, and confidence in knowledge-based cognitive tasks aided by an artificially intelligent aid.

RRMSE-enhanced weighted voting regressor for improved ensemble regression.

Ensemble regression methods are widely used to improve prediction accuracy by combining multiple regression models, especially when dealing with continuous numerical targets. However, most ensemble voting regressors use equal weights for each base model's predictions, which can limit their effectiveness, particularly when there is no specific domain knowledge to guide the weighting. This uniform weighting approach doesn't consider that some models may perform better than others on different datasets, leaving room for improvement in optimizing ensemble performance. To overcome this limitation, we propose the RRMSE (Relative Root Mean Square Error) Voting Regressor, a new ensemble regression technique that assigns weights to each base model based on their relative error rates. By using an RRMSE-based weighting function, our method gives more importance to models that demonstrate higher accuracy, thereby enhancing the overall prediction quality. We tested the RRMSE Voting Regressor on six popular regression datasets and compared its performance with several state-of-the-art ensemble regression algorithms. The results show that the RRMSE Voting Regressor consistently achieves lower prediction errors than existing methods across all tested datasets. This improvement highlights the effectiveness of using relative error metrics for weighting in ensemble models. Our approach not only fills a gap in current ensemble regression techniques but also provides a reliable and adaptable method for boosting prediction performance in various machine learning tasks. By leveraging the strengths of individual models through smart weighting, the RRMSE Voting Regressor offers a significant advancement in the field of ensemble learning.

Explainable Artificial Intelligence to Quantify Adenoid Hypertrophy-related Upper Airway Obstruction using 3D Shape Analysis.

Explainable Artificial Intelligence to Quantify Adenoid Hypertrophy-related Upper Airway Obstruction using 3D Shape Analysis.

What a shot! Effect of biathlon and badminton on attention of preteens in middle schools: a longitudinal study

ABSTRACT The aim of the present study was to test the effect of a biathlon and badminton programme, two activities requiring high cognitive engagement, on attention, executive functions, and visuo-manual coordination of preteens in a school context. We compared these two intervention programmes with a running programme involving aerobic fitness but less cognitive engagement. We hypothesised that badminton and biathlon would improve preteens’ attention compared with running. The intervention programmes consisted of ten 90-minute physical education sessions carried out by middle school teachers and were randomly assigned to each class (badminton, N = 60; biathlon, N = 81; running, N = 52). Multivariate analysis revealed no significant effects on overall performance. However, an exploratory analysis accounting for individual differences in each cognitive measure indicated improved performance in divided attention and Stroop tasks for participants in the biathlon and badminton interventions, with the latter influenced by speed performance. Additionally, action video games were found to moderate the relationship between interventions and performance in the visuo-manual coordination task. For the first time, biathlon activity, which combines running and shooting, was tested in a curricular context and suggested potential benefits for attention, executive functions, and visuo-manual coordination. These findings underscore the importance of developing and experimentally validating evidence-based pedagogical content in physical education while fostering collaboration with teachers and education professionals.

Bayesian Random Forest with Multiple Imputation by Chain Equations for High-Dimensional Missing Data: A Simulation Study

The pervasive challenge of missing data in scientific research forces a critical trade-off: discarding incomplete observations, which risks significant information loss, while conventional imputation methods struggle to maintain accuracy in high-dimensional settings. Although approaches like multiple imputation (MI) and random forest (RF) proximity-based imputation offer improvements over naive deletion, they exhibit limitations in complex missing data scenarios or sparse high-dimensional settings. To address these gaps, we propose a novel integration of Multiple Imputation by Chained Equations (MICE) with Bayesian Random Forest (BRF), leveraging MICE’s iterative flexibility and BRF’s probabilistic robustness to enhance the imputation accuracy and downstream predictive performance. Our hybrid framework, BRF-MICE, uniquely combines the efficiency of MICE’s chained equations with BRF’s ability to quantify uncertainty through Bayesian tree ensembles, providing stable parameter estimates even under extreme missingness. We empirically validate this approach using synthetic datasets with controlled missingness mechanisms (MCAR, MAR, MNAR) and dimensionality, contrasting it against established methods, including RF and Bayesian Additive Regression Trees (BART). The results demonstrate that BRF-MICE achieves a superior performance in classification and regression tasks, with a 15–20% lower error under varying missingness conditions compared to RF and BART while maintaining computational scalability. The method’s iterative Bayesian updates effectively propagate imputation uncertainty, reducing overconfidence in high-dimensional predictions, a key weakness of frequentist alternatives.

Discriminator guided visible-to-infrared image translation

This paper proposes a discriminator-guided visible-to-infrared image translation algorithm based on a generative adversarial network and designs a multi-scale fusion generative network. The generative network enhances the perception of the image’s fine-grained features by fusing features of different scales in the channel direction. Meanwhile, the discriminator performs the infrared image reconstruction task, which provides additional infrared information to train the generator. This enhances the convergence efficiency of generator training through soft label guidance generated through knowledge distillation. The experimental results show that compared to the existing typical infrared image generation algorithms, the proposed method can generate higher-quality infrared images and achieve better performance in both subjective visual description and objective metric evaluation, and that it has better performance in the downstream tasks of the template matching and image fusion tasks.

Gait variability predicts post-fatigue obstacle course performance among military cadets: An exploratory study.

Gait variability predicts post-fatigue obstacle course performance among military cadets: An exploratory study.

Active Supervised Cross-Modal Retrieval.

Supervised Cross-Modal Retrieval (SCMR) achieves significant performance with the supervision provided by substantial label annotations of multi-modal data. However, the requirement for large annotated multi-modal datasets restricts the use of supervised cross-modal retrieval in many practical scenarios. Active Learning (AL) has been proposed to reduce labeling costs while improving performance in various label-dependent tasks, in which the most informative unlabeled samples are selected for labeling and training. Directly exploiting the existing AL methods for supervised cross-modal retrieval may not be a good idea since they only focus on the uncertainty within each modality, ignoring the inter-modality relationship within the text-image pairs. Furthermore, existing methods focus exclusively on the informativeness of data during sample selection, leading to a biased, homogenized set where selected samples often contain nearly identical semantics and are densely distributed in a region of the feature space. Persistent training with such biased data selections can disturb multi-modal representation learning and substantially degrade the retrieval performance of SCMR. In this work, we propose an Active Supervised Cross-Modal Retrieval (ASCMR) framework, which effectively identifies informative multi-modal samples and generates unbiased sample selections. In particular, we propose a probabilistic multi-modal informativeness estimation that captures both the intra-modality and inter-modality uncertainty of multi-modal pairs within a unified representation. To ensure unbiased sample selection, we introduce a density-aware budget allocation strategy that constrains the active learning objective of maximizing the informativeness of selection with a novel semantic density regularization term. The proposed methods are evaluated on three widely used benchmark datasets, MS-COCO, NUS-WIDE, and MIRFlickr, demonstrating our effectiveness in significantly reducing the annotation cost while outperforming other baselines of active learning strategies. We could achieve over 95% of the fully supervised model's performance by only utilizing 6%, 3%, and 4% active selected samples for MS-COCO, NUS-WIDE, and MIRFlickr, respectively. Our source code is available at https://github.com/openimmc/ascmr.

YOLOv8-ADDA: an enhanced algorithm for real-time traffic sign detection

Abstract Existing deep learning-based traffic sign recognition methods often face challenges such as low accuracy, slow detection response, and limited applicability in real-world scenarios. This paper proposes an improved model based on YOLOv8, aiming to enhance detection accuracy and reduce model complexity. To achieve this, the Adaptive Downsampling (ADown) method is integrated into the Backbone Conv module, effectively reducing input feature dimensions and minimizing redundant parameters. Additionally, the Deformable Attention Transformer (DAT) mechanism is incorporated into the C2f module, dynamically adjusting the attention range to capture more relevant visual features. Furthermore, Wise and Efficient IoU (Wise-EIoU) is employed as the loss function to improve classification performance in object detection tasks. The proposed algorithm is evaluated on the Roadsign and CCTSDB2021 datasets, achieving significant performance improvements. On the Roadsign dataset, the model achieves a precision of 94.9%, with mAP50 and mAP50:95 reaching 97.1% and 85.9%, respectively, surpassing the YOLOv8n benchmark by 2.7% and 4.2% on the same metrics. The number of model parameters is reduced to 2.79M, reflecting a 7.5% decrease compared to YOLOv8n. To verify its practical applicability, experiments are conducted on the PIX micro smart car platform. The proposed model achieves an inference speed of 59 FPS, representing a 15.6% improvement over the baseline, demonstrating its suitability for real-time traffic sign detection tasks. These results validate the effectiveness of the YOLOv8-ADDA model in improving detection accuracy and reducing computational complexity.

Dual-Domain Triple Contrast for Cross-Dataset Skeleton-Based Action Recognition

Skeleton-based Action Recognition (SAR) is widely recognized for its robustness and efficiency in human action analysis, but its performance in cross-dataset tasks has been limited due to domain shifts between different datasets. To address this challenge, current methods typically approach cross-dataset SAR as an Unsupervised Domain Adaptation (UDA) task, which is tackled using domain adaptation or self-supervised learning strategies. In this article, we propose a Dual-Domain Triple Contrast (D2TC) framework for cross-dataset SAR under the UDA setting. Unlike existing UDA methods that either focus on a single strategy or superficially combine strategies, our D2TC leverages contrastive learning to integrate both strategies into a unified framework. It performs three types of contrastive learning: Self-Supervised Contrastive Learning, Supervised Contrastive Learning, and UDA with Contrastive Learning, across both source and target domains. The triple contrasts go beyond mere summation, effectively bridging the domain gap and enhancing the model’s representational capacity. Additionally, we introduce multi-modal ensemble contrast and extreme skeleton augmentation methods to further enhance the skeleton-based representation learning. Extensive experiments on six cross-dataset settings validate the superiority of our D2TC framework over state-of-the-art methods, demonstrating its effectiveness in reducing domain discrepancies and improving cross-dataset SAR performance. The codes are available on https://github.com/KennCoder7/DualDomainTripleContrast .

ANF-Net: A Refined Segmentation Network for Road Scenes with Multiple Noises and Various Morphologies of Cracks

Cracks are a common early road defect that tends to worsen with the aging of roads, potentially leading to severe structural damage. Timely and accurate crack detection plays a crucial role in mitigating such risks and holds significant importance for infrastructure maintenance. Deep learning techniques have demonstrated excellent performance in image-based crack extraction tasks. However, challenges persist due to the presence of numerous noisy pixels in the image background and the diverse and intricate morphologies of cracks, leading to issues such as misclassification and omission. To address these issues, this paper proposes a refined pixel-level segmentation network (ANF-Net) suitable for complex crack detection scenarios with high noise levels and diverse crack morphologies. When extracting crack features, on one hand, the network introduces an attention module tailored for crack scenes to learn pixel-wise feature weights, enabling the network to focus on crack regions and thereby reducing the impact of similar background features, mitigating false positives caused by noise misclassification. On the other hand, a constrained multi-morphological convolution structure is constructed by imposing learnable continuous constraints on the deformation offsets of convolutional kernels, allowing the network to adaptively fit different crack shapes. This design enhances the network’s ability to extract cracks in morphologically diverse, narrow, and densely populated regions, effectively preventing issues such as crack extraction interruptions and omissions. Additionally, a multi-scale discrete wavelet transform enhancement module is designed to assist the network in considering frequency domain information that contains crack features, further improving its feature extraction capability. Simulations are conducted using three publicly available crack datasets, and the proposed method is compared with mainstream segmentation models. The results demonstrate that the proposed method achieves F1 scores of 87.9%, 82.5%, and 71.5% on the three datasets, respectively, all of which surpass the performance of current mainstream segmentation models. The proposed network accurately extracts road cracks and exhibits robust performance.

A Lightweight TA-YOLOv8 Method for the Spot Weld Surface Anomaly Detection of Body in White

The deep learning architecture YOLO (You Only Look Once) has demonstrated its superior visual detection performance in various computer vision tasks and has been widely applied in the field of automatic surface defect detection. In this paper, we propose a lightweight YOLOv8-based method for the quality inspection of car body welding spots. We developed a TA-YOLOv8 network structure which has an improved Task-Aligned (TA) head detection, designed to handle a small sample size, imbalanced positive and negative samples, and high-noise characteristics of Body-in-White welding spot data. By learning with fewer parameters, the model achieves more efficient and accurate classification. Additionally, our algorithm framework can perform anomaly segmentation and classification on our open-world raw datasets obtained from actual production environments. The experimental results show that the lightweight module improves the processing speed by an average of 2.8%, with increases in detection the mAP@50-95 and recall rate of 1.35% and 0.1226, respectively.

FEGGNN: Feature-Enhanced Gated Graph Neural Network for robust few-shot skin disease classification.

FEGGNN: Feature-Enhanced Gated Graph Neural Network for robust few-shot skin disease classification.

TCM-KLLaMA: Intelligent generation model for Traditional Chinese Medicine Prescriptions based on knowledge graph and large language model.

TCM-KLLaMA: Intelligent generation model for Traditional Chinese Medicine Prescriptions based on knowledge graph and large language model.

Evaluating the contribution of parallel processing of color and shape in a conjunction search task

Traditionally, researchers interpret the difficulty in conjunction search as difficulty in binding features. In the present study, we used a behavioral-computational approach to assess if parameters from feature search could predict performance in a color-shape conjunction search task. We also investigated whether pooling-mediated processing in peripheral regions was a key-limiting factor in performance in conjunction search by manipulating display arrangements across different experiments. The results indicated that parameters in homogeneous search displays can indeed be used to successfully predict performance in conjunction search displays. This finding is noteworthy because it indicates that the visual system must be extracting the same information from the display in feature and conjunction search tasks (i.e., the target-distractor similarity) using color and shape. Furthermore, there was no compelling evidence that pooling-mediated processing was the primary constraint on performance in this conjunction search task. A model-comparison approach compared the accuracy of different distractor rejection architectures in predicting performance in conjunction search tasks. The winning model showed participants engaging hierarchically with the display, selecting and rejecting distractor subsets based on a single defining feature. Taken in the context of previous research on heterogeneous search performance, the current results imply that the inherent demands of search for a conjunction of color and shape compel participants to adopt this targeted search strategy.

FusionESP: Improved Enzyme-Substrate Pair Prediction by Fusing Protein and Chemical Knowledge.

To reduce the cost of the experimental characterization of the potential substrates for enzymes, machine learning prediction models offer an alternative solution. Pretrained language models, as powerful approaches for protein and molecule representation, have been employed in the development of enzyme-substrate prediction models, achieving promising performance. In addition to continuing improvements in language models, effectively fusing encoders to handle multimodal prediction tasks is critical for further enhancing model performance by using available representation methods. Here, we present FusionESP, a multimodal architecture that integrates protein and chemistry language models with two independent projection heads and a contrastive learning strategy for predicting enzyme-substrate pairs. Our best model achieved state-of-the-art performance with an accuracy of 94.77% on independent test data and exhibited better generalization capacity while requiring fewer computational resources and training data, compared to previous studies of a fine-tuned encoder or employing more encoders. It also confirmed our hypothesis that embeddings of positive pairs are closer to each other in a high-dimension space, while negative pairs exhibit the opposite trend. Our ablation studies showed that the projection heads played a crucial role in performance enhancement, while the contrastive learning strategy further improved the projection heads' capacity in classification tasks. The proposed architecture is expected to be further applied to enhance performance in additional multimodality prediction tasks in biology. A user-friendly web server of FusionESP is established and freely accessible at https://rqkjkgpsyu.us-east-1.awsapprunner.com/.

A dual-stream feature decomposition network with weight transformation for multi-modality image fusion

As an image enhancement technology, multi-modal image fusion primarily aims to retain salient information from multi-source image pairs in a single image, generating imaging information that contains complementary features and can facilitate downstream visual tasks. However, dual-stream methods with convolutional neural networks (CNNs) as backbone networks predominantly have limited receptive fields, whereas methods with Transformers are time-consuming, and both lack the exploration of cross-domain information. This study proposes an innovative image fusion model designed for multi-modal images, encompassing pairs of infrared and visible images and multi-source medical images. Our model leverages the strengths of both Transformers and CNNs to model various feature types effectively, addressing both short- and long-range learning as well as the extraction of low- and high-frequency features. First, our shared encoder is constructed based on Transformers for long-range learning, including an intra-modal feature extraction block, an inter-modal feature extraction block, and a novel feature alignment block that handles slight misalignments. Our private encoder for extracting low- and high-frequency features employs a dual-stream architecture based on CNNs, which includes a dual-domain selection mechanism and an invertible neural network. Second, we develop a cross-attention-based Swin Transformer block to explore cross-domain information. In particular, we introduce a weight transformation that is embedded into the Transformer block to enhance the efficiency. Third, a unified loss function incorporating a dynamic weighting factor is formulated to capture the inherent commonalities of multi-modal images. A comprehensive qualitative and quantitative analysis of image fusion and object detection experimental results demonstrates that the proposed method effectively preserves thermal targets and background texture details, surpassing state-of-the-art alternatives in terms of achieving high-quality image fusion and improving the performance in subsequent visual tasks.

Transcutaneous and transcranial electrical stimulation for enhancing military performance: an update and systematic review.

Electrical stimulation (ES), including transcranial electrical stimulation (tES) and transcutaneous vagus nerve stimulation (tVNS), has shown potential for cognitive enhancement in military contexts. Various types of ES, such as transcranial direct current stimulation (tDCS) and transcranial alternating current stimulation (tACS), modulate neuronal membrane potentials and cortical excitability, potentially improving cognitive functions relevant to military training and operations. This systematic review updates previous findings by examining studies published between 2019 and 2024 that investigated electrical stimulation effects on cognitive performance in military personnel and tasks. We focused on whether the studies addressed key questions about the generalizability of lab findings to military tasks, the frequency and intensity of adverse effects, the impact of repeated ES administration, and the ethical and regulatory considerations for its use in potentially vulnerable military populations. Eleven studies met the inclusion criteria; most demonstrated overall low to some concerns, however, two of these had overall high risk of bias. While tES and tVNS showed some promise for enhancing multitasking and visual search performance, the results were mixed, with no reliable effects on vigilance tasks. The reviewed studies highlight the need for a better understanding of ES mechanisms, optimal stimulation parameters, and individual differences in response to ES. They also highlight the importance of conducting high-powered research in military settings to evaluate the efficacy, safety, and ethical implications of ES. Future research should address the generalizability of lab-based results to real-world military tasks, monitor the frequency and intensity of adverse effects, and explore the long-term impacts of repeated administration. Furthermore, ethical and regulatory considerations are crucial for the responsible application of ES in military contexts, and a series of outstanding questions is posed to guide continuing research in this domain.