Search

The acyl-CoA-binding domain-containing protein 6 (ACBD6) is ubiquitously expressed, plays a role in the acylation of lipids and proteins and regulates the N-myristoylation of proteins via N-myristoyltransferase enzymes (NMTs). However, its precise function in cells is still unclear, as is the consequence of ACBD6 defects on human pathophysiology. Using exome sequencing and extensive international data sharing efforts, we identified 45 affected individuals from 28 unrelated families (consanguinity 93%) with bi-allelic pathogenic, predominantly loss-of-function (18/20) variants in ACBD6. We generated zebrafish and Xenopus tropicalis acbd6 knockouts by CRISPR/Cas9 and characterized the role of ACBD6 on protein N-myristoylation with myristic acid alkyne (YnMyr) chemical proteomics in the model organisms and human cells, with the latter also being subjected further to ACBD6 peroxisomal localization studies. The affected individuals (23 males and 22 females), aged 1-50 years, typically present with a complex and progressive disease involving moderate-to-severe global developmental delay/intellectual disability (100%) with significant expressive language impairment (98%), movement disorders (97%), facial dysmorphism (95%) and mild cerebellar ataxia (85%) associated with gait impairment (94%), limb spasticity/hypertonia (76%), oculomotor (71%) and behavioural abnormalities (65%), overweight (59%), microcephaly (39%) and epilepsy (33%). The most conspicuous and common movement disorder was dystonia (94%), frequently leading to early-onset progressive postural deformities (97%), limb dystonia (55%) and cervical dystonia (31%). A jerky tremor in the upper limbs (63%), a mild head tremor (59%), parkinsonism/hypokinesia developing with advancing age (32%) and simple motor and vocal tics were among other frequent movement disorders. Midline brain malformations including corpus callosum abnormalities (70%), hypoplasia/agenesis of the anterior commissure (66%), short midbrain an

Diabetic Retinopathy (DR) stands as the leading cause of blindness globally, particularly affecting individuals between the ages of 20 and 70. This paper presents a Computer-Aided Diagnosis (CAD) system designed for the automatic classification of retinal images into five distinct classes: Normal, Mild, Moderate, Severe, and Proliferative Diabetic Retinopathy (PDR). The proposed system leverages Convolutional Neural Networks (CNNs) employing pre-trained deep learning models. Through the application of fine-tuning techniques, our model is trained on fundus images of diabetic retinopathy with resolutions of 350x350x3 and 224x224x3. Experimental results obtained on the Kaggle platform, utilizing resources comprising 4 CPUs, 17 GB RAM, and 1 GB Disk, demonstrate the efficacy of our approach. The achieved Area Under the Curve (AUC) values for CNN, MobileNet, VGG-16, InceptionV3, and InceptionResNetV2 models are 0.50, 0.70, 0.53, 0.63, and 0.69, respectively., Comment: 3 pages, 1 figure, 1 table

Face inpainting, the technique of restoring missing or damaged regions in facial images, is pivotal for applications like face recognition in occluded scenarios and image analysis with poor-quality captures. This process not only needs to produce realistic visuals but also preserve individual identity characteristics. The aim of this paper is to inpaint a face given periocular region (eyes-to-face) through a proposed new Generative Adversarial Network (GAN)-based model called Eyes-to-Face Network (E2F-Net). The proposed approach extracts identity and non-identity features from the periocular region using two dedicated encoders have been used. The extracted features are then mapped to the latent space of a pre-trained StyleGAN generator to benefit from its state-of-the-art performance and its rich, diverse and expressive latent space without any additional training. We further improve the StyleGAN output to find the optimal code in the latent space using a new optimization for GAN inversion technique. Our E2F-Net requires a minimum training process reducing the computational complexity as a secondary benefit. Through extensive experiments, we show that our method successfully reconstructs the whole face with high quality, surpassing current techniques, despite significantly less training and supervision efforts. We have generated seven eyes-to-face datasets based on well-known public face datasets for training and verifying our proposed methods. The code and datasets are publicly available.

This paper explores the application of large language models (LLMs), like ChatGPT, for biometric tasks. We specifically examine the capabilities of ChatGPT in performing biometric-related tasks, with an emphasis on face recognition, gender detection, and age estimation. Since biometrics are considered as sensitive information, ChatGPT avoids answering direct prompts, and thus we crafted a prompting strategy to bypass its safeguard and evaluate the capabilities for biometrics tasks. Our study reveals that ChatGPT recognizes facial identities and differentiates between two facial images with considerable accuracy. Additionally, experimental results demonstrate remarkable performance in gender detection and reasonable accuracy for the age estimation tasks. Our findings shed light on the promising potentials in the application of LLMs and foundation models for biometrics.

Self-supervised learning (SSL) approaches have achieved great success when the amount of labeled data is limited. Within SSL, models learn robust feature representations by solving pretext tasks. One such pretext task is contrastive learning, which involves forming pairs of similar and dissimilar input samples, guiding the model to distinguish between them. In this work, we investigate the application of contrastive learning to the domain of medical image analysis. Our findings reveal that MoCo v2, a state-of-the-art contrastive learning method, encounters dimensional collapse when applied to medical images. This is attributed to the high degree of inter-image similarity shared between the medical images. To address this, we propose two key contributions: local feature learning and feature decorrelation. Local feature learning improves the ability of the model to focus on the local regions of the image, while feature decorrelation removes the linear dependence among the features. Our experimental findings demonstrate that our contributions significantly enhance the model's performance in the downstream task of medical segmentation, both in the linear evaluation and full fine-tuning settings. This work illustrates the importance of effectively adapting SSL techniques to the characteristics of medical imaging tasks. The source code will be made publicly available at: https://github.com/CAMMA-public/med-moco, Comment: Accepted at at ISBI-2024 (https://biomedicalimaging.org/2024/). 4 pages, 2 figures, 2 tables

Vision transformers are known to be more computationally and data-intensive than CNN models. These transformer models such as ViT, require all the input image tokens to learn the relationship among them. However, many of these tokens are not informative and may contain irrelevant information such as unrelated background or unimportant scenery. These tokens are overlooked by the multi-head self-attention (MHSA), resulting in many redundant and unnecessary computations in MHSA and the feed-forward network (FFN). In this work, we propose a method to optimize the amount of unnecessary interactions between unimportant tokens by separating and sending them through a different low-cost computational path. Our method does not add any parameters to the ViT model and aims to find the best trade-off between training throughput and achieving a 0% loss in the Top-1 accuracy of the final model. Our experimental results on training ViT-small from scratch show that SkipViT is capable of effectively dropping 55% of the tokens while gaining more than 13% training throughput and maintaining classification accuracy at the level of the baseline model on Huawei Ascend910A.

Breast cancer is the most common malignancy affecting women worldwide and is notable for its morphologic and biologic diversity, with varying risks of recurrence following treatment. The Oncotype DX Breast Recurrence Score test is an important predictive and prognostic genomic assay for estrogen receptor-positive breast cancer that guides therapeutic strategies; however, such tests can be expensive, delay care, and are not widely available. The aim of this study was to develop a multi-model approach integrating the analysis of whole slide images and clinicopathologic data to predict their associated breast cancer recurrence risks and categorize these patients into two risk groups according to the predicted score: low and high risk. The proposed novel methodology uses convolutional neural networks for feature extraction and vision transformers for contextual aggregation, complemented by a logistic regression model that analyzes clinicopathologic data for classification into two risk categories. This method was trained and tested on 993 hematoxylin and eosin-stained whole-slide images of breast cancers with corresponding clinicopathological features that had prior Oncotype DX testing. The model's performance was evaluated using an internal test set of 198 patients from Dartmouth Health and an external test set of 418 patients from the University of Chicago. The multi-model approach achieved an AUC of 0.92 (95 percent CI: 0.88-0.96) on the internal set and an AUC of 0.85 (95 percent CI: 0.79-0.90) on the external cohort. These results suggest that with further validation, the proposed methodology could provide an alternative to assist clinicians in personalizing treatment for breast cancer patients and potentially improving their outcomes., Comment: 16 pages, 4 figures and 4 tables

To address the massive growth of data traffic over cellular networks, increasing spatial reuse of the frequency spectrum by the deployment of small base stations (SBSs) has been considered. For rapid deployment of SBSs in the networks, caching popular content along with new coded caching schemes are proposed. To maximize the cellular network's capacity, densifying it with small base stations is inevitable. In ultra-dense cellular networks, coverage of SBSs may overlap. To this aim, the multi-access caching system, where users potentially can access multiple cache nodes simultaneously, has attracted more attention in recent years. Most previous works on multi-access coded caching, only consider specific conditions such as cyclic wrap-around network topologies. In this paper, we investigate caching in ultra-dense cellular networks, where different users can access different numbers of caches under non-uniform content popularity distribution, and propose Multi-Access Hybrid coded-uncoded Caching (MAHC). We formulate the optimization problem of the proposed scheme for general network topologies and evaluate it for 2-SBS network scenarios. The numerical and simulation results show that the proposed MAHC scheme outperforms optimal conventional uncoded and previous multi-access coded caching (MACC) schemes., Comment: 10 pages

Despite the recent remarkable achievement in gaze estimation, efficient and accurate personalization of gaze estimation without labels is a practical problem but rarely touched on in the literature. To achieve efficient personalization, we take inspiration from the recent advances in Natural Language Processing (NLP) by updating a negligible number of parameters, "prompts", at the test time. Specifically, the prompt is additionally attached without perturbing original network and can contain less than 1% of a ResNet-18's parameters. Our experiments show high efficiency of the prompt tuning approach. The proposed one can be 10 times faster in terms of adaptation speed than the methods compared. However, it is non-trivial to update the prompt for personalized gaze estimation without labels. At the test time, it is essential to ensure that the minimizing of particular unsupervised loss leads to the goals of minimizing gaze estimation error. To address this difficulty, we propose to meta-learn the prompt to ensure that its updates align with the goal. Our experiments show that the meta-learned prompt can be effectively adapted even with a simple symmetry loss. In addition, we experiment on four cross-dataset validations to show the remarkable advantages of the proposed method. Code is available at https://github.com/hmarkamcan/TPGaze., Comment: Accepted by AAAI 2024

The rapid expansion of computational workloads in contemporary computing necessitates continuous advancements in computational systems. The proliferation of many-core architectures, with a doubling core count every two months, poses challenges in inter-core communication and data mobility. These challenges, driven by limitations in bandwidth and physical interconnections, impede the scalability of supercomputing systems. Research endeavors have been exploring strategies to optimize inter-core communications, including wireless integration and innovative broadcast mechanisms, to enhance the performance of high-performance computing (HPC). Many-core systems play a pivotal role in HPC, enabling concurrent task execution and unprecedented parallelism. This work seeks to improve HPC performance and scalability, focusing on inter-core communication in chiplet-based architectures. Traditional wired links in multi-chiplet systems face scalability and performance limitations, including intricate wiring, hardware failure risks, latency, and bandwidth challenges. The advent of wireless Network-on-Chip (NoC) technology offers an innovative solution, leveraging wireless communication to overcome these limitations. Wireless NoC provides benefits such as reduced complexity, flexible chip design, dynamic reconfiguration, and mitigating latency issues. However, the adoption of wireless NoC introduces unique challenges like interference, security, and power management. This thesis conducts a comprehensive exploration of wireless NoC's potential and challenges, aiming to advance high-performance computing and set the stage for future advancements. The research further introduces a hybrid approach in chiplet-based architectures for MPI collectives, where wired and wireless links are strategically combined. This approach optimizes message distribution, resulting in a substantial performance improvement. The findings emphasize an improved system speedup. In summary, this work underscores

Mental manipulation, a significant form of abuse in interpersonal conversations, presents a challenge to identify due to its context-dependent and often subtle nature. The detection of manipulative language is essential for protecting potential victims, yet the field of Natural Language Processing (NLP) currently faces a scarcity of resources and research on this topic. Our study addresses this gap by introducing a new dataset, named ${\rm M{\small ental}M{\small anip}}$, which consists of $4,000$ annotated movie dialogues. This dataset enables a comprehensive analysis of mental manipulation, pinpointing both the techniques utilized for manipulation and the vulnerabilities targeted in victims. Our research further explores the effectiveness of leading-edge models in recognizing manipulative dialogue and its components through a series of experiments with various configurations. The results demonstrate that these models inadequately identify and categorize manipulative content. Attempts to improve their performance by fine-tuning with existing datasets on mental health and toxicity have not overcome these limitations. We anticipate that ${\rm M{\small ental}M{\small anip}}$ will stimulate further research, leading to progress in both understanding and mitigating the impact of mental manipulation in conversations., Comment: Accepted at ACL 2024

Machine learning models have been an inevitable tool for analyzing medical images by radiologists. These models provide important information about the contents of these images using extracted radiomic features. However, the dimensionality of the feature space can cause reduction in the accuracy of prediction, a phenomenon known as the curse of dimensionality. In this study we propose a feature selection method using an autoencoder, which incorporates the performance of a classifier within the feature selection process. This is achieved by automatically adjusting a threshold value used for selecting the features fed to the classifier. The contribution of this study is twofold. The first contribution is an improvement to group lasso to include the group size as a cost parameter of the autoencoder. The second contribution is to automate the selection of the threshold value used for eliminating redundant input features. The threshold value in our proposed method is learned during training phase of the proposed model. Our experimental results indicates that the proposed model can successfully converge to appropriate feature selection parameters.

Membranes are ubiquitous tools for modern water treatment technology that critically eliminate hazardous materials such as organic, inorganic, heavy metals, and biomedical pollutants. Nowadays, nano-membranes are of particular interest for myriad applications such as water treatment, desalination, ion exchange, ion concentration control, and several kinds of biomedical applications. However, this state-of-the-art technology suffers from some drawbacks, e.g., toxicity and fouling of contaminants, which makes the synthesis of green and sustainable membranes indeed safety-threatening. Typically, sustainability, non-toxicity, performance optimization, and commercialization are concerns centered on manufacturing green synthesized membranes. Thus, critical issues related to toxicity, biosafety, and mechanistic aspects of green-synthesized nano-membranes have to be systematically and comprehensively reviewed and discussed. Herein we evaluate various aspects of green nano-membranes in terms of their synthesis, characterization, recycling, and commercialization aspects. Nanomaterials intended for nano-membrane development are classified in view of their chemistry/synthesis, advantages, and limitations. Indeed, attaining prominent adsorption capacity and selectivity in green-synthesized nano-membranes requires multi-objective optimization of a number of materials and manufacturing parameters. In addition, the efficacy and removal performance of green nano-membranes are analyzed theoretically and experimentally to provide researchers and manufacturers with a comprehensive image of green nano-membrane efficiency under real environmental conditions., Validerad;2023;Nivå 2;2023-06-26 (sofila)

Machine learning models have been an inevitable tool for analyzing medical images by radiologists. These models provide important information about the contents of these images using extracted radiomic features. However, the dimensionality of the feature space can cause reduction in the accuracy of prediction, a phenomenon known as the curse of dimensionality. In this study we propose a feature selection method using an autoencoder, which incorporates the performance of a classifier within the feature selection process. This is achieved by automatically adjusting a threshold value used for selecting the features fed to the classifier. The contribution of this study is twofold. The first contribution is an improvement to group lasso to include the group size as a cost parameter of the autoencoder. The second contribution is to automate the selection of the threshold value used for eliminating redundant input features. The threshold value in our proposed method is learned during training phase of the proposed model. Our experimental results indicates that the proposed model can successfully converge to appropriate feature selection parameters.

Regulating the local configuration of atomically dispersed transition-metal atom catalysts is the key to oxygen electrocatalysis performance enhancement. Unlike the previously reported single-atom or dual-atom configurations, we designed a new type of binary-atom catalyst, through engineering Fe-N4 electronic structure with adjacent Co-N2C2 and nitrogen-coordinated Co nanoclusters, as oxygen electrocatalysts. The resultant optimized electronic structure of the Fe-N4 active center favors the binding capability of intermediates and enhances oxygen reduction reaction (ORR) activity in both alkaline and acid conditions. In addition, anchoring M–N–C atomic sites on highly graphitized carbon supports guarantees of efficient charge- and mass-transports, and escorts the high bifunctional catalytic activity of the entire catalyst. Further, through the combination of electrochemical studies and in-situ X-ray absorption spectroscopy analyses, the ORR degradation mechanisms under highly oxidative conditions during oxygen evolution reaction processes were revealed. This work developed a new binary-atom catalyst and systematically investigates the effect of highly oxidative environments on ORR electrochemical behavior. It demonstrates the strategy for facilitating oxygen electrocatalytic activity and stability of the atomically dispersed M–N–C catalysts.

Background: Colorectal cancer (CRC) is one of the most common cancers in the world. Some dietary factors such as fat intake have been identified as the risk factors for CRC. This study aimed to investigate the effect of fat mass and obesity-associated (FTO) gene rs9939609 polymorphism on the association between CRC and different types of dietary fats. Methods: This case-control study was performed on 135 CRC cases and 294 healthy controls in Tehran, Iran. Data on demographic factors, anthropometric measurements, physical activity, the intake of different types of dietary fats, and FTO gene rs9939609 polymorphism was collected from all participants. The association between cancer and dietary fat intake in individuals with different FTO genotypes was assessed using different models of logistic regression. Results: Oleic acid intake was higher in the case group compared to the control group in both people with TT (7.2±3.46 vs. 5.83±3.06 g/d, P=0.02) and AA/AT genotypes (8.7±6.23 vs. 5.57 ±3.2 g/d, P<0.001). Among carriers of AA/AT genotypes of FTO rs9939609 polymorphism, a positive association was found between CRC and higher intakes of oleic acid (OR=1.12, CI95% 1.03–1.21, P=0.01) and cholesterol (OR=1.01, CI95% 1.00–1.02; P=0.01) after adjusting for age, sex, physical activity, alcohol use, smoking, calorie intake, and body mass index. Conclusion: Higher intakes of cholesterol and oleic acid were associated with a higher risk of CRC in FTO-risk allele carriers. The association of CRC and dietary fat may be influenced by the FTO genotype. Further longitudinal studies are warranted to confirm these findings.

Background: Colorectal cancer (CRC) is one of the most common cancers in the world. Some dietary factors such as fat intake have been identified as the risk factors for CRC. This study aimed to investigate the effect of fat mass and obesity-associated (FTO) gene rs9939609 polymorphism on the association between CRC and different types of dietary fats. Methods: This case-control study was performed on 135 CRC cases and 294 healthy controls in Tehran, Iran. Data on demographic factors, anthropometric measurements, physical activity, the intake of different types of dietary fats, and FTO gene rs9939609 polymorphism was collected from all participants. The association between cancer and dietary fat intake in individuals with different FTO genotypes was assessed using different models of logistic regression. Results: Oleic acid intake was higher in the case group compared to the control group in both people with TT (7.2±3.46 vs. 5.83±3.06 g/d, P=0.02) and AA/AT genotypes (8.7±6.23 vs. 5.57 ±3.2 g/d, P<0.001). Among carriers of AA/AT genotypes of FTO rs9939609 polymorphism, a positive association was found between CRC and higher intakes of oleic acid (OR=1.12, CI95% 1.03–1.21, P=0.01) and cholesterol (OR=1.01, CI95% 1.00–1.02; P=0.01) after adjusting for age, sex, physical activity, alcohol use, smoking, calorie intake, and body mass index. Conclusion: Higher intakes of cholesterol and oleic acid were associated with a higher risk of CRC in FTO-risk allele carriers. The association of CRC and dietary fat may be influenced by the FTO genotype. Further longitudinal studies are warranted to confirm these findings.

Background: Colorectal cancer (CRC) is one of the most common cancers in the world. Some dietary factors such as fat intake have been identified as the risk factors for CRC. This study aimed to investigate the effect of fat mass and obesity-associated (FTO) gene rs9939609 polymorphism on the association between CRC and different types of dietary fats. Methods: This case-control study was performed on 135 CRC cases and 294 healthy controls in Tehran, Iran. Data on demographic factors, anthropometric measurements, physical activity, the intake of different types of dietary fats, and FTO gene rs9939609 polymorphism was collected from all participants. The association between cancer and dietary fat intake in individuals with different FTO genotypes was assessed using different models of logistic regression. Results: Oleic acid intake was higher in the case group compared to the control group in both people with TT (7.2±3.46 vs. 5.83±3.06 g/d, P=0.02) and AA/AT genotypes (8.7±6.23 vs. 5.57 ±3.2 g/d, P<0.001). Among carriers of AA/AT genotypes of FTO rs9939609 polymorphism, a positive association was found between CRC and higher intakes of oleic acid (OR=1.12, CI95% 1.03–1.21, P=0.01) and cholesterol (OR=1.01, CI95% 1.00–1.02; P=0.01) after adjusting for age, sex, physical activity, alcohol use, smoking, calorie intake, and body mass index. Conclusion: Higher intakes of cholesterol and oleic acid were associated with a higher risk of CRC in FTO-risk allele carriers. The association of CRC and dietary fat may be influenced by the FTO genotype. Further longitudinal studies are warranted to confirm these findings.

Membranes are ubiquitous tools for modern water treatment technology that critically eliminate hazardous materials such as organic, inorganic, heavy metals, and biomedical pollutants. Nowadays, nano-membranes are of particular interest for myriad applications such as water treatment, desalination, ion exchange, ion concentration control, and several kinds of biomedical applications. However, this state-of-the-art technology suffers from some drawbacks, e.g., toxicity and fouling of contaminants, which makes the synthesis of green and sustainable membranes indeed safety-threatening. Typically, sustainability, non-toxicity, performance optimization, and commercialization are concerns centered on manufacturing green synthesized membranes. Thus, critical issues related to toxicity, biosafety, and mechanistic aspects of green-synthesized nano-membranes have to be systematically and comprehensively reviewed and discussed. Herein we evaluate various aspects of green nano-membranes in terms of their synthesis, characterization, recycling, and commercialization aspects. Nanomaterials intended for nano-membrane development are classified in view of their chemistry/synthesis, advantages, and limitations. Indeed, attaining prominent adsorption capacity and selectivity in green-synthesized nano-membranes requires multi-objective optimization of a number of materials and manufacturing parameters. In addition, the efficacy and removal performance of green nano-membranes are analyzed theoretically and experimentally to provide researchers and manufacturers with a comprehensive image of green nano-membrane efficiency under real environmental conditions., Validerad;2023;Nivå 2;2023-06-26 (sofila)

The small amount of training data for many state-of-the-art deep learning-based Face Recognition (FR) systems causes a marked deterioration in their performance. Although a considerable amount of research has addressed this issue by inventing new data augmentation techniques, using either input space transformations or Generative Adversarial Networks (GAN) for feature space augmentations, these techniques have yet to satisfy expectations. In this paper, we propose an approach named the Face Representation Augmentation (FRA) for augmenting face datasets. To the best of our knowledge, FRA is the first method that shifts its focus towards manipulating the face embeddings generated by any face representation learning algorithm to create new embeddings representing the same identity and facial emotion but with an altered posture. Extensive experiments conducted in this study convince of the efficacy of our methodology and its power to provide noiseless, completely new facial representations to improve the training procedure of any FR algorithm. Therefore, FRA can help the recent state-of-the-art FR methods by providing more data for training FR systems. The proposed method, using experiments conducted on the Karolinska Directed Emotional Faces (KDEF) dataset, improves the identity classification accuracies by 9.52 %, 10.04 %, and 16.60 %, in comparison with the base models of MagFace, ArcFace, and CosFace, respectively.

Image segmentation is a branch of computer vision that is widely used in real world applications including biomedical image processing. With recent advancement of deep learning, image segmentation has achieved at a very high level performance. Recently, UNet architecture is found as the core of novel deep learning segmentation methods. In this paper we implement UNet architecture from scratch with using basic blocks in Pytorch and evaluate its performance on multiple biomedical image datasets. We also use transfer learning to apply novel modified UNet segmentation packages on the biomedical image datasets. We fine tune the pre-trained transferred model with each specific dataset. We compare its performance with our fundamental UNet implementation. We show that transferred learning model has better performance in image segmentation than UNet model that is implemented from scratch.

In digital pathology, whole slide images (WSIs) are widely used for applications such as cancer diagnosis and prognosis prediction. Visual transformer models have recently emerged as a promising method for encoding large regions of WSIs while preserving spatial relationships among patches. However, due to the large number of model parameters and limited labeled data, applying transformer models to WSIs remains challenging. Inspired by masked language models, we propose a pretext task for training the transformer model without labeled data to address this problem. Our model, MaskHIT, uses the transformer output to reconstruct masked patches and learn representative histological features based on their positions and visual features. The experimental results demonstrate that MaskHIT surpasses various multiple instance learning approaches by 3% and 2% on survival prediction and cancer subtype classification tasks, respectively. Furthermore, MaskHIT also outperforms two of the most recent state-of-the-art transformer-based methods. Finally, a comparison between the attention maps generated by the MaskHIT model with pathologist's annotations indicates that the model can accurately identify clinically relevant histological structures in each task.

In recent years, breeding programs have increased significantly in size and complexity, with various highly interdependent parameters and many contrasting breeding goals. As a result, resource allocation in these programs has become more complex, and deriving an optimal breeding strategy has become increasingly challenging. To address this, a common practice is to reduce the optimization problem to a set of scenarios that differ only in a few parameters and can therefore be analyzed in detail. The goal of this article is to provide a framework for the numerical optimization of breeding programs that goes beyond the simple comparison of scenarios. For this, we first determine the space of potential breeding programs only limited by basic constraints like the budget and housing capacities. Subsequently, the goal is to identify the optimal breeding program by finding the parametrization that maximizes the target function by combining different breeding goals. To assess the value of the target function for a parametrization, we propose using stochastic simulations and the subsequent use of a kernel regression method to cope with the stochasticity of simulation outcomes. This procedure is performed iteratively to narrow down the most promising areas of the search space and perform more and more simulations in these areas of interest. In a simplified example applied to a dairy cattle program, our proposed framework has shown its ability to identify an optimal breeding strategy that aligns with a target function aiming at genetic gain and genetic diversity conservation limited by budget constraints.

A problem with machine learning (ML) techniques for detecting intrusions in the Internet of Things (IoT) is that they are ineffective in the detection of low-frequency intrusions. In addition, as ML models are trained using specific attack categories, they cannot recognize unknown attacks. This article integrates strategies of cost-sensitive learning and multitask learning into a hybrid ML model to address these two challenges. The hybrid model consists of an autoencoder for feature extraction and a support vector machine (SVM) for detecting intrusions. In the cost-sensitive learning phase for the class imbalance problem, the hinge loss layer is enhanced to make a classifier strong against low-distributed intrusions. Moreover, to detect unknown attacks, we formulate the SVM as a multitask problem. Experiments on the UNSW-NB15 and BoT-IoT datasets demonstrate the superiority of our model in terms of recall, precision, and F1-score averagely 92.2%, 96.2%, and 94.3%, respectively, over other approaches., Green Open Access added to TU Delft Institutional Repository ‘You share, we take care!’ – Taverne project https://www.openaccess.nl/en/you-share-we-take-care Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

A number of tasks have been proposed recently to facilitate easy access to charts such as chart QA and summarization. The dominant paradigm to solve these tasks has been to fine-tune a pretrained model on the task data. However, this approach is not only expensive but also not generalizable to unseen tasks. On the other hand, large language models (LLMs) have shown impressive generalization capabilities to unseen tasks with zero- or few-shot prompting. However, their application to chart-related tasks is not trivial as these tasks typically involve considering not only the underlying data but also the visual features in the chart image. We propose PromptChart, a multimodal few-shot prompting framework with LLMs for chart-related applications. By analyzing the tasks carefully, we have come up with a set of prompting guidelines for each task to elicit the best few-shot performance from LLMs. We further propose a strategy to inject visual information into the prompts. Our experiments on three different chart-related information consumption tasks show that with properly designed prompts LLMs can excel on the benchmarks, achieving state-of-the-art., Comment: 23 pages

Endometrial cancer, the fourth most common cancer in females in the United States, with the lifetime risk for developing this disease is approximately 2.8% in women. Precise histologic evaluation and molecular classification of endometrial cancer is important for effective patient management and determining the best treatment modalities. This study introduces EndoNet, which uses convolutional neural networks for extracting histologic features and a vision transformer for aggregating these features and classifying slides based on their visual characteristics into high- and low- grade. The model was trained on 929 digitized hematoxylin and eosin-stained whole-slide images of endometrial cancer from hysterectomy cases at Dartmouth-Health. It classifies these slides into low-grade (Endometroid Grades 1 and 2) and high-grade (endometroid carcinoma FIGO grade 3, uterine serous carcinoma, carcinosarcoma) categories. EndoNet was evaluated on an internal test set of 110 patients and an external test set of 100 patients from the public TCGA database. The model achieved a weighted average F1-score of 0.91 (95% CI: 0.86-0.95) and an AUC of 0.95 (95% CI: 0.89-0.99) on the internal test, and 0.86 (95% CI: 0.80-0.94) for F1-score and 0.86 (95% CI: 0.75-0.93) for AUC on the external test. Pending further validation, EndoNet has the potential to support pathologists without the need of manual annotations in classifying the grades of gynecologic pathology tumors., Comment: 4 Tables and 3 Figures

Over the last few decades, researchers have been focusing on determining the critical attack angle at which dynamic stall occurs. This angle is usually determined by solving the Navier-Stokes equations, which include viscosity, pressure, gravity, and acceleration terms. However, Navier-Stokes equations are quite complex to solve and consequently difficult to simulate, thus the simulation is not accurate enough. Therefore, this article predicts the critical attack angle for the first time using Euler equations devoid of viscous terms. One of the key advantages of Euler equations is their ability to capture the vortices and predict stall dynamics. The Euler equations are thus integrated and the resulting equations are discretized using the finite volume method. A first-order upwind-based method is used to calculate the convective fluxes at the cell boundaries in the finite volume approach. A NACA 0012 airfoil is chosen for this study at various attack angles with a Mach number of 0.3. Based on the justification of Crocco's theorem, the Euler equations successfully act as Navier-Stokes equations. The vortex patterns are found to behave independently of the artificial dissipation. All the vortices are successfully predicted using the inviscid governing equations. The numerical results obtained are validated by other published experimental and numerical data.

This paper presents an extensive exploration and comparative analysis of lightweight face recognition (FR) models, specifically focusing on MobileFaceNet and its modified variant, MMobileFaceNet. The need for efficient FR models on devices with limited computational resources has led to the development of models with reduced memory footprints and computational demands without sacrificing accuracy. Our research delves into the impact of dataset selection, model architecture, and optimization algorithms on the performance of FR models. We highlight our participation in the EFaR-2023 competition, where our models showcased exceptional performance, particularly in categories restricted by the number of parameters. By employing a subset of the Webface42M dataset and integrating sharpness-aware minimization (SAM) optimization, we achieved significant improvements in accuracy across various benchmarks, including those that test for cross-pose, cross-age, and cross-ethnicity performance. The results underscore the efficacy of our approach in crafting models that are not only computationally efficient but also maintain high accuracy in diverse conditions.

In this paper, we present SwiftLearn, a data-efficient approach to accelerate training of deep learning models using a subset of data samples selected during the warm-up stages of training. This subset is selected based on an importance criteria measured over the entire dataset during warm-up stages, aiming to preserve the model performance with fewer examples during the rest of training. The importance measure we propose could be updated during training every once in a while, to make sure that all of the data samples have a chance to return to the training loop if they show a higher importance. The model architecture is unchanged but since the number of data samples controls the number of forward and backward passes during training, we can reduce the training time by reducing the number of training samples used in each epoch of training. Experimental results on a variety of CV and NLP models during both pretraining and finetuning show that the model performance could be preserved while achieving a significant speed-up during training. More specifically, BERT finetuning on GLUE benchmark shows that almost 90% of the data can be dropped achieving an end-to-end average speedup of 3.36x while keeping the average accuracy drop less than 0.92%.

Massive transformer-based models face several challenges, including slow and computationally intensive pre-training and over-parametrization. This paper addresses these challenges by proposing a versatile method called GQKVA, which generalizes query, key, and value grouping techniques. GQKVA is designed to speed up transformer pre-training while reducing the model size. Our experiments with various GQKVA variants highlight a clear trade-off between performance and model size, allowing for customized choices based on resource and time limitations. Our findings also indicate that the conventional multi-head attention approach is not always the best choice, as there are lighter and faster alternatives available. We tested our method on ViT, which achieved an approximate 0.3% increase in accuracy while reducing the model size by about 4% in the task of image classification. Additionally, our most aggressive model reduction experiment resulted in a reduction of approximately 15% in model size, with only around a 1% drop in accuracy.

Large Language Models (LLMs) have significantly advanced the field of Natural Language Processing (NLP), but their lack of interpretability has been a major concern. Current methods for interpreting LLMs are post hoc, applied after inference time, and have limitations such as their focus on low-level features and lack of explainability at higher level text units. In this work, we introduce proto-lm, a prototypical network-based white-box framework that allows LLMs to learn immediately interpretable embeddings during the fine-tuning stage while maintaining competitive performance. Our method's applicability and interpretability are demonstrated through experiments on a wide range of NLP tasks, and our results indicate a new possibility of creating interpretable models without sacrificing performance. This novel approach to interpretability in LLMs can pave the way for more interpretable models without the need to sacrifice performance., Comment: Accepted to the Findings of EMNLP 2023

Recent advances in whole-slide image (WSI) scanners and computational capabilities have significantly propelled the application of artificial intelligence in histopathology slide analysis. While these strides are promising, current supervised learning approaches for WSI analysis come with the challenge of exhaustively labeling high-resolution slides - a process that is both labor-intensive and time-consuming. In contrast, self-supervised learning (SSL) pretraining strategies are emerging as a viable alternative, given that they don't rely on explicit data annotations. These SSL strategies are quickly bridging the performance disparity with their supervised counterparts. In this context, we introduce an SSL framework. This framework aims for transferable representation learning and semantically meaningful clustering by synergizing invariance loss and clustering loss in WSI analysis. Notably, our approach outperforms common SSL methods in downstream classification and clustering tasks, as evidenced by tests on the Camelyon16 and a pancreatic cancer dataset., Comment: Accepted by ICCV2023

This paper presents the summary of the Efficient Face Recognition Competition (EFaR) held at the 2023 International Joint Conference on Biometrics (IJCB 2023). The competition received 17 submissions from 6 different teams. To drive further development of efficient face recognition models, the submitted solutions are ranked based on a weighted score of the achieved verification accuracies on a diverse set of benchmarks, as well as the deployability given by the number of floating-point operations and model size. The evaluation of submissions is extended to bias, cross-quality, and large-scale recognition benchmarks. Overall, the paper gives an overview of the achieved performance values of the submitted solutions as well as a diverse set of baselines. The submitted solutions use small, efficient network architectures to reduce the computational cost, some solutions apply model quantization. An outlook on possible techniques that are underrepresented in current solutions is given as well., Comment: Accepted at IJCB 2023

Purpose: To present a 65-year-old patient with macular hole (MH) and submacular hemorrhage (SMH) secondary to a ruptured retinal arterial macroaneurysm (RAM) which was successfully treated with a novel surgical technique.Patient: A 65-year-old woman presented with a 1-week history of sudden-onset visual loss in her right eye. Her best-corrected visual acuity (BCVA) was hand motion in the right eye. Her fundus examination revealed intraretinal and massive subretinal hemorrhage with macular involvement. An MH with an approximate diameter of 600 microns was also noted. The patient was treated by a standard pars plana vitrectomy (PPV) and internal limiting membrane (ILM) peeling, followed by 10 µg recombinant tissue plasminogen activator injection into the submacular space via the MH with a 25-gauge vitrectomy probe. BCVA of the patient increased to 20/320 and 20/60, one and four weeks after surgery, respectively. Optical coherence tomography (OCT) images confirmed the complete anatomical closure of the MH and the restoration of the outer retinal layers such as the external limiting membrane and the ellipsoid zone.Conclusion: This case report expands our knowledge about the management of MH in combination with SMH after a ruptured RAM. We suggest the use of a vitrectomy probe and MH for subretinal recombinant tissue plasminogen activator (rtPA) injection.

Sludge is a nutrient-rich organic waste generated from wastewater treatment plants. However, the application of sludge as a nutrient source is limited by its high contents of water and pollutants. In this study, the effects of biomass type on nutrient recovery and heavy metal removal from digested sludge by hydrothermal treatment (HTT) were investigated. Blending biomass with digested sludge for HTT at 180–240 °C increased the recovery of nitrogen in the treated solids. At the HTT temperature of 240 °C, HTT with hardwood sawdust led to the highest nitrogen recovery of 70.6%, compared to the lowest nitrogen recovery of 36.5% without biomass. Blending biomass slightly decreased the recovery of phosphorus compared to those without biomass. Nevertheless, the lowest phosphorus recovery of 91.3% with the use of hardwood sawdust at the HTT temperature of 240 °C was only ∼7.0% less than that without biomass. Blending biomass reduced the contents of macro-metals such as Ca, Fe, Mg and Al in treated solids but the metal contents varied with different biomasses. Regarding the heavy metals, the use of rice husk did not decrease the contents of Ni and Co while blending bagasse did not decrease the content of Cr at HTT temperatures of 210 °C and 240 °C compared to the use of other biomasses. The different effects of biomass type on nutrient recovery and heavy metals were likely related to the types and abundances of organic acids such as acetic acid, oxygen-containing functional groups such as C–OH and COOH, oxide minerals such as silica from biomasses and the overall effects of these factors. This study provides very useful information in selection of lignocellulosic biomass for HTT of sludge for nutrient recovery and heavy metal removal.

It has been desirable to create biodegradable polymer materials that are thermally stable, flame- retardant and mechanically robust. One promising strategy is to develop fire retardant polylactide (PLA) using lignin as inexpensive fillers, but existing functionalized lignin cannot improve the thermal stability of PLA, in addition to unsatisfactory fire retardancy. Herein, we report a lignin-derived fire retardant (LF) by grafting polyphosphoramide into the chemical structure of lignin waste through graft polymerization. The results show that with 8 wt% of LF, the final PLA shows significantly enhanced thermal stability in N2. Moreover, the resulting PLA achieves a satisfactory UL-94 V-0 rating, meeting the demanding requirement of fire retardancy in industries. Additionally, the PLA/LF composite displays increased elastic modulus and well-preserved strength and ductility of pure PLA. This work provides an effective strategy for creating biobased fire retardant additives using industrial lignin waste, and PLA biocomposites which hold great promise for packaging, fabrics, electronic, and automotive applications.

In this paper, we will use the categorical approach to Hilbert (C^{ast})-modules over a commutative (C^{ast})-algebra to investigate the approximately orthogonality preserving mappings on Hilbert (C^{ast})-modules over a commutative (C^{ast})-algebra. Indeed, we show that if (Psi:Gamma rightarrow Gamma^{prime} ) is a nonzero ( C_{0}(Z) )-linear (( delta , varepsilon))-orthogonality preserving mapping between the continuous fields of Hilbert spaces on a locally compact Hausdorff space (Z), then (Psi) is injective, continuous and also for every ( x, y in Gamma ) and (z in Z), [ vert langle Psi(x),Psi(y) rangle(z) - varphi^2(z) langle x,y rangle(z) vert leq frac{4(varepsilon - delta)}{(1-delta)(1+varepsilon)} Vert Psi(x) Vert Vert Psi(y) Vert, ] where (varphi(z) = sup { Vert Psi(u)(z) Vert : u ~ text{is a unit vector in} ~ Gamma }).

Redox potential (Eh) control plays a significant role during sulfide mineral flotation by influencing the reactions on the surface of the minerals and accordingly the flotation behaviour. In this study, the metallurgical performance of typical copper sulfide minerals, molybdenite as well as gangue minerals (e.g., pyrite, tennantite, and enargite) under different pH and Eh conditions of the flotation cell were investigated. The copper and molybdenum processing plant at the Sungun complex-Iran were selected as a case study. For this purpose, Eh of flotation cells of phases 1 and 2 of copper and molybdenum processing circuits – Sungun complex – were measured by off-line method. After performing chemical analysis, the mineralogical study of the input load and products of each of the aforementioned flotation circuits in the rougher, cleaner, re-cleaner, and scavenger stages was performed. Based on the results, the potential in cells of phases 1 and 2 of copper concentration plants is in the range of −60 to −100 mV; and for the molybdenum plant, is in the range of −500 to −700 mV. The potentials of more than −100 mV in the phases of copper concentration plants have created suitable conditions for the separation of copper sulfide and molybdenite minerals from gangue minerals, especially pyrite. Adjustment of Eh in the range of −500 to −700 mV in the molybdenum processing plant has also led to the depression of copper minerals and the flotation of molybdenite, resulting in the effective separation of these minerals. However, grade analysis and mineralogical studies indicate the misplaced copper minerals into tailings, the passage of chalcopyrite and pyrite to molybdenum concentrate, the misplaced molybdenite to copper concentrate, and also the presence of minerals containing harmful elements such as arsenic in copper concentrate. Eh fluctuations in phase 1 and 2 of copper plants, the interaction of copper sulfide minerals, especially chalcopyrite with pyrite (and the depres, Validerad;2022;Nivå 2;2022-03-18 (sofila)

Redox potential (Eh) control plays a significant role during sulfide mineral flotation by influencing the reactions on the surface of the minerals and accordingly the flotation behaviour. In this study, the metallurgical performance of typical copper sulfide minerals, molybdenite as well as gangue minerals (e.g., pyrite, tennantite, and enargite) under different pH and Eh conditions of the flotation cell were investigated. The copper and molybdenum processing plant at the Sungun complex-Iran were selected as a case study. For this purpose, Eh of flotation cells of phases 1 and 2 of copper and molybdenum processing circuits – Sungun complex – were measured by off-line method. After performing chemical analysis, the mineralogical study of the input load and products of each of the aforementioned flotation circuits in the rougher, cleaner, re-cleaner, and scavenger stages was performed. Based on the results, the potential in cells of phases 1 and 2 of copper concentration plants is in the range of −60 to −100 mV; and for the molybdenum plant, is in the range of −500 to −700 mV. The potentials of more than −100 mV in the phases of copper concentration plants have created suitable conditions for the separation of copper sulfide and molybdenite minerals from gangue minerals, especially pyrite. Adjustment of Eh in the range of −500 to −700 mV in the molybdenum processing plant has also led to the depression of copper minerals and the flotation of molybdenite, resulting in the effective separation of these minerals. However, grade analysis and mineralogical studies indicate the misplaced copper minerals into tailings, the passage of chalcopyrite and pyrite to molybdenum concentrate, the misplaced molybdenite to copper concentrate, and also the presence of minerals containing harmful elements such as arsenic in copper concentrate. Eh fluctuations in phase 1 and 2 of copper plants, the interaction of copper sulfide minerals, especially chalcopyrite with pyrite (and the depres, Validerad;2022;Nivå 2;2022-03-18 (sofila)

Data-intensive workloads and applications, such as machine learning (ML), are fundamentally limited by traditional computing systems based on the von-Neumann architecture. As data movement operations and energy consumption become key bottlenecks in the design of computing systems, the interest in unconventional approaches such as Near-Data Processing (NDP), machine learning, and especially neural network (NN)-based accelerators has grown significantly. Emerging memory technologies, such as ReRAM and 3D-stacked, are promising for efficiently architecting NDP-based accelerators for NN due to their capabilities to work as both high-density/low-energy storage and in/near-memory computation/search engine. In this paper, we present a survey of techniques for designing NDP architectures for NN. By classifying the techniques based on the memory technology employed, we underscore their similarities and differences. Finally, we discuss open challenges and future perspectives that need to be explored in order to improve and extend the adoption of NDP architectures for future computing platforms. This paper will be valuable for computer architects, chip designers, and researchers in the area of machine learning., This work has been supported by the CoCoUnit ERC Advanced Grant of the EU’s Horizon 2020 program (grant No 833057), the Spanish State Research Agency (MCIN/AEI) under grant PID2020-113172RB-I00, and the ICREA Academia program., Peer Reviewed, Postprint (published version)

Habrobracon hebetor Say is an important biological control agent that has a wide range of lepidopteran hosts, especially from the Noctuidae and Pyralidae families. Nutrition is a very important element in the efficiency of natural enemies in integrated pest management (IPM). In the presented research, the sublethal effects of five selected chemical and botanical insecticides, fenvalerate, propargite, buprofezin, dayabon and palizin, on the activity of amylase and protease enzymes in the midgut of this important ectoparasitoid wasp were studied under laboratory conditions. Changes in the activity of enzymes were investigated by LC30 treatment of the female wasps, which were 0.14, 7.01, 3.04, 9.58 and   2.53 mg a.i./mL, respectively. The obtained results showed that this parasitoid wasp has a long midgut in its digestive tract. Also, there were significant differences among the enzymatic treatments (F5,12 =14.695 and 99.278) except protein concentration (F5,12 = 0.670). Regarding the amylase enzyme, the highest and lowest activities were obtained in the control and fenvalerate treatment (0.0435±0.0023 mU/mg vs. 0.0277±0.0010 mU/mg). In addition, the highest proteolytic activity was observed in the control (4.9817±0.0268 U/mg) and the lowest in the propargite treatment (3.2231±0.0917 U/mg). By closely investigating the results, dayabon and palizin showed the lowest negative effects on the enzymes and could be applied together with this parasitoid wasp in IPM designs.

Deep learning has been effective for histology image analysis in digital pathology. However, many current deep learning approaches require large, strongly- or weakly-labeled images and regions of interest, which can be time-consuming and resource-intensive to obtain. To address this challenge, we present HistoPerm, a view generation method for representation learning using joint embedding architectures that enhances representation learning for histology images. HistoPerm permutes augmented views of patches extracted from whole-slide histology images to improve classification performance. We evaluated the effectiveness of HistoPerm on two histology image datasets for Celiac disease and Renal Cell Carcinoma, using three widely used joint embedding architecture-based representation learning methods: BYOL, SimCLR, and VICReg. Our results show that HistoPerm consistently improves patch- and slide-level classification performance in terms of accuracy, F1-score, and AUC. Specifically, for patch-level classification accuracy on the Celiac disease dataset, HistoPerm boosts BYOL and VICReg by 8% and SimCLR by 3%. On the Renal Cell Carcinoma dataset, patch-level classification accuracy is increased by 2% for BYOL and VICReg, and by 1% for SimCLR. In addition, on the Celiac disease dataset, models with HistoPerm outperform the fully-supervised baseline model by 6%, 5%, and 2% for BYOL, SimCLR, and VICReg, respectively. For the Renal Cell Carcinoma dataset, HistoPerm lowers the classification accuracy gap for the models up to 10% relative to the fully-supervised baseline. These findings suggest that HistoPerm can be a valuable tool for improving representation learning of histopathology features when access to labeled data is limited and can lead to whole-slide classification results that are comparable to or superior to fully-supervised methods.

Edge caching (EC) decreases the average access delay of the end-users through caching popular content at the edge network, however, it increases the leakage probability of valuable information such as users preferences. Most of the existing privacy-preserving approaches focus on adding layers of encryption, which confronts the network with more challenges such as energy and computation limitations. We employ a chunk-based joint probabilistic caching (JPC) approach to mislead an adversary eavesdropping on the communication inside an EC and maximizing the adversary's error in estimating the requested file and requesting cache. In JPC, we optimize the probability of each cache placement to minimize the communication cost while guaranteeing the desired privacy and then, formulate the optimization problem as a linear programming (LP) problem. Since JPC inherits the curse of dimensionality, we also propose scalable JPC (SPC), which reduces the number of feasible cache placements by dividing files into non-overlapping subsets. We also compare the JPC and SPC approaches against an existing probabilistic method, referred to as disjoint probabilistic caching (DPC) and random dummy-based approach (RDA). Results obtained through extensive numerical evaluations confirm the validity of the analytical approach, the superiority of JPC and SPC over DPC and RDA.

Machine learning (ML) models have been applied to a wide range of natural language processing (NLP) tasks in recent years. In addition to making accurate decisions, the necessity of understanding how models make their decisions has become apparent in many applications. To that end, many interpretability methods that help explain the decision processes of ML models have been developed. Yet, there currently exists no widely-accepted metric to evaluate the quality of explanations generated by these methods. As a result, there currently is no standard way of measuring to what degree an interpretability method achieves an intended objective. Moreover, there is no accepted standard of performance by which we can compare and rank the current existing interpretability methods. In this paper, we propose a novel metric for quantifying the quality of explanations generated by interpretability methods. We compute the metric on three NLP tasks using six interpretability methods and present our results.