This paper describes the system used by the Machine Learning Group of LTU in subtask 1 of the SemEval-2022 Task 4: Patronizing and Condescending Language (PCL) Detection. Our system consists of finetuning a pretrained text-to-text transfer transformer (T5) and innovatively reducing its out-of-class predictions. The main contributions of this paper are 1) the description of the implementation details of the T5 model we used, 2) analysis of the successes & struggles of the model in this task, and 3) ablation studies beyond the official submission to ascertain the relative importance of data split. Our model achieves an F1 score of 0.5452 on the official test set.

In this paper, we first describe various synchronous and asynchronous methods for enhancing student engagement in big online courses. We showcase the implementation of these methods in the “Introduction to Artificial Intelligence (AI)” course at Luleå University of Technology, which has attracted around 500 students in each of its iterations (twice yearly, since 2019). We also show that these methods can be applied efficiently, in terms of the teaching hours required. With the increase in digitization and student mobility, the demand for improved and personalized content delivery for distance education has also increased. This applies not only in the context of traditional undergraduate education, but also in the context of adult education and lifelong learning. This higher level of demand, however, introduces a challenge, especially as it is typically combined with a shortage of staff and needs for efficient education. This challenge is further amplified by the current pandemic situation, which led to an even bigger risk of student-dropout. To mitigate this risk, as well as to meet the increased demand, we applied various methods for creating engaging interaction in our pedagogy based on Moor’s framework: learner-to-learner, learner-to-instructor, and learner-to-content engagement strategies. The main methods of this pedagogy are as follows: short, and interactive videos, active discussions in topic-based forums, regular live sessions with group discussions, and the introduction of optional content at many points in the course, to address different target groups. In this paper, we show how we originally designed and continuously improved the course, without requiring more than 500 teaching hours per iteration (one hour per enrolled student), while we also managed to increase the successful completion rate of the participants by 10%, and improved student engagement and feedback for the course by 50%. We intend to share a set of best-practices applicable to many other e-learning courses in ICT.

Image contrast enhancement refers to an operation of remapping the pixels’ values of an image to emphasize desired information in the image. In this work, we propose a novel pixel-based (local) contrast enhancement algorithm, based on the human visual perception. First, we make an observation that pixels with lower regional contrast should be amplified for the purpose of enhancing the contrast and pixels with higher regional contrast should be suppressed to avoid undesired over-enhancement. To determine the quality of the regional contrast in the image (either lower or higher), a reference image will be created using a proposed global based contrast enhancement method (termed as Mean Brightness Bidirectional Histogram Equalization in the paper) for fast computation reason. To quantify the abovementioned regional contrast, we propose a method based on human visual perception taking Just Noticeable Difference (JND) into account. In short, our proposed algorithm is able to limit the enhancement of well-contrasted regions and enhance the poor contrast regions in an image. Both objective quality and subjective quality experimental results suggested that the proposed algorithm enhances images consistently across images with different dynamic range. We conclude that the proposed algorithm exhibits excellent consistency in producing satisfactory result for different type of images. It is important to note that the algorithm can be directly implemented in color space and not limited only to grayscale. The proposed algorithm can be obtained from the following GitHub link: https://github.com/UTARSL1/CHE.

“He who knows no foreign languages knows nothing of his own.” Johann Wolfgang emphasized the worth of languages for expanding ones learning horizons. This work instills the same notion into the industrial internet of things (IIoT) sensory devices paradigm. We study the interoperability problem setting with a new perspective of envisioning knowledge graphs (KGs) modeling for the device to device ontology alignment. Ontology alignment is structured as entity alignment in which similar entities are linked from two heterogeneous knowledge graphs. The novelty is conceiving the IIoT ontology graph as a language of the sensory device and then addressing it through the natural language processing (NLP) language translation approach. The IIoT ontology graph nodes have unique URIs so they act as words (sentences) for the NLP model and the schema of the graph is depicted as the language structure. Existing methods give less attention to the importance of structural information which ignores the fact of even when a node pair has similar entity labels it may not refer to a similar context and vice versa. To deal with these issues, we propose a novel solution using a modified BERT_INT model on graph Triplets for ontology alignment among heterogeneous IIoT devices. Moreover, an iterative framework is designed to leverage the alignments within nodes as well as among relations. As the first attempt at this problem, the proposed model is tested on a contemporary language dataset of DBP15K and compared with the best state-of-the-art results. The proposed model outperforms the target baseline BERT_INT model by 2.1% in terms of HR@1, HR@10, and MRR. Next, a dataset on ontology instances is constructed on smart building sensors using two W3C standardized IIoT ontologies i.e. SSN and SOSA. Comprehensive experiments and analysis with ablation study on language and structural encoders demonstrate the effectiveness of our model.

The technical capabilities of modern Industry 4.0 and Industry 5.0 are vast and growing exponentially daily. The present-day Industrial Internet of Things (IIoT) combines manifold underlying technologies that require real-time interconnection and communication among heterogeneous devices. Smart cities are established with sophisticated designs and control of seamless machine-to-machine (M2M) communication, to optimize resources, costs, performance, and energy distributions. All the sensory devices within a building interact to maintain a sustainable climate for residents and intuitively optimize the energy distribution to optimize energy production. However, this encompasses quite a few challenges for devices that lack a compatible and interoperable design. The conventional solutions are restricted to limited domains or rely on engineers designing and deploying translators for each pair of ontologies. This is a costly process in terms of engineering effort and computational resources. An issue persists that a new device with a different ontology must be integrated into an existing IoT network. We propose a self-learning model that can determine the taxonomy of devices given their ontological meta-data and structural information. The model finds matches between two distinct ontologies using a natural language processing (NLP) approach to learn linguistic contexts. Then, by visualizing the ontological network as a knowledge graph, it is possible to learn the structure of the meta-data and understand the device's message formulation. Finally, the model can align entities of ontological graphs that are similar in context and structure.Furthermore, the model performs dynamic M2M translation without requiring extra engineering or hardware resources.

In the modern manufacturing industry, collaborative architectures are growing in popularity. We propose an Industry 5.0 value-driven manufacturing process automation ecosystem in which each edge automation system is based on a local cloud and has a service-oriented architecture. Additionally, we integrate cloud-based collaborative learning (CCL) across building energy management, logistic robot management, production line management, and human worker Aide local clouds to facilitate shared learning and collaborate in generating manufacturing workflows. Consequently, the workflow management system generates the most effective and Industry 5.0-driven workflow recipes. In addition to managing energy for a sustainable climate and executing a cost-effective, optimized, and resilient manufacturing process, this work ensures the well-being of human workers. This work has significant implications for future work, as the ecosystem can be deployed and tested for any industrial use case.

The fourth and fifth industrial revolutions (Industry 4.0 and Industry 5.0) have driven significant advances in digitalization and integration of advanced technologies, emphasizing the need for sustainable solutions. Smart Energy Systems (SESs) have emerged as crucial tools for addressing climate change, integrating smart grids and smart homes/buildings to improve energy infrastructure. To achieve a robust and sustainable SES, stakeholders must collaborate efficiently through an energy management framework based on the Internet of Things (IoT). Demand Response (DR) is key to balancing energy demands and costs. This research proposes an edge-based automation cloud solution, utilizing Eclipse Arrowhead local clouds, which are based on Service-Oriented Architecture that promotes the integration of stakeholders. This novel solution guarantees secure, low-latency communication among various smart home and industrial IoT technologies. The study also introduces a theoretical framework that employs AI at the edge to create environment profiles for smart buildings, optimizing DR and ensuring human comfort. By focusing on room-level optimization, the research aims to improve the overall efficiency of SESs and foster sustainable energy practices.

Document classification is one of the most critical steps in the document analysis pipeline. There are two types of approaches for document classification, known as image-based and multimodal approaches. Image-based document classification approaches are solely based on the inherent visual cues of the document images. In contrast, the multimodal approach co-learns the visual and textual features, and it has proved to be more effective. Nonetheless, these approaches require a huge amount of data. This paper presents a novel approach for document classification that works with a small amount of data and outperforms other approaches. The proposed approach incorporates a hierarchical attention network (HAN) for the textual stream and the EfficientNet-B0 for the image stream. The hierarchical attention network in the textual stream uses dynamic word embedding through fine-tuned BERT. HAN incorporates both the word level and sentence level features. While earlier approaches rely on training on a large corpus (RVL-CDIP), we show that our approach works with a small amount of data (Tobacco-3482). To this end, we trained the neural network at Tobacco-3482 from scratch. Therefore, we outperform the state-of-the-art by obtaining an accuracy of 90.3%. This results in a relative error reduction rate of 7.9%.

This comprehensive systematic review critically analyzes the current progress and challenges in automating transtibial prosthesis alignment. The manual identification of alignment changes in prostheses has been found to lack reliability, necessitating the development of automated processes. Through a rigorous systematic search across major electronic databases, this review includes the highly relevant studies out of an initial pool of 2111 records. The findings highlight the urgent need for automated alignment systems in individuals with transtibial amputation. The selected studies represent cutting-edge research, employing diverse approaches such as advanced machine learning algorithms and innovative alignment tools, to automate the detection and adjustment of prosthesis alignment. Collectively, this review emphasizes the immense potential of automated transtibial prosthesis alignment systems to enhance alignment accuracy and significantly reduce human error. Furthermore, it identifies important limitations in the reviewed studies, serving as a catalyst for future research to address these gaps and explore alternative machine learning algorithms. The insights derived from this systematic review provide valuable guidance for researchers, clinicians, and developers aiming to propel the field of automated transtibial prosthesis alignment forward.

Depth maps produced by LiDAR-based approaches are sparse. Even high-end LiDAR sensors produce highly sparse depth maps, which are also noisy around the object boundaries. Depth completion is the task of generating a dense depth map from a sparse depth map. While the earlier approaches focused on directly completing this sparsity from the sparse depth maps, modern techniques use RGB images as a guidance tool to resolve this problem. Whilst many others rely on affinity matrices for depth completion. Based on these approaches, we have divided the literature into two major categories; unguided methods and image-guided methods. The latter is further subdivided into multi-branch and spatial propagation networks. The multi-branch networks further have a sub-category named image-guided filtering. In this paper, for the first time ever we present a comprehensive survey of depth completion methods. We present a novel taxonomy of depth completion approaches, review in detail different state-of-the-art techniques within each category for depth completion of LiDAR data, and provide quantitative results for the approaches on KITTI and NYUv2 depth completion benchmark datasets.

The ongoing COVID-19 pandemic has brought online education to the forefront of pedagogical discussions. To make this increased interest sustainable in a post-pandemic era, online courses must be built on strong pedagogical foundations. With a long history of pedagogic research, there are many principles, frameworks, and models available to help teachers in doing so. These models cover different teaching perspectives, such as constructive alignment, feedback, and the learning environment. In this paper, we discuss how we designed and implemented our online Natural Language Processing (NLP) course following constructive alignment and adhering to the pedagogical principles of LTU. By examining our course and analyzing student evaluation forms, we show that we have met our goal and successfully delivered the course. Furthermore, we discuss the additional benefits resulting from the current mode of delivery, including the increased reusability of course content and increased potential for collaboration between universities. Lastly, we also discuss where we can and will further improve the current course design.

Handwritten signatures in biometric authentication leverage unique individual characteristics for identification, offering high specificity through dynamic and static properties. However, this modality faces significant challenges from sophisticated forgery attempts, underscoring the need for enhanced security measures in common applications. To address forgery in signature-based biometric systems, integrating a forgery-resistant modality, namely, noninvasive electroencephalography (EEG), which captures unique brain activity patterns, can significantly enhance system robustness by leveraging multimodality’s strengths. By combining EEG, a physiological modality, with handwritten signatures, a behavioral modality, our approach capitalizes on the strengths of both, significantly fortifying the robustness of biometric systems through this multimodal integration. In addition, EEG’s resistance to replication offers a high-security level, making it a robust addition to user identification and verification. This study presents a new multimodal SignEEG v1.0 dataset based on EEG and hand-drawn signatures from 70 subjects. EEG signals and hand-drawn signatures have been collected with Emotiv Insight and Wacom One sensors, respectively. The multimodal data consists of three paradigms based on mental, & motor imagery, and physical execution: i) thinking of the signature’s image, (ii) drawing the signature mentally, and (iii) drawing a signature physically. Extensive experiments have been conducted to establish a baseline with machine learning classifiers. The results demonstrate that multimodality in biometric systems significantly enhances robustness, achieving high reliability even with limited sample sizes. We release the raw, pre-processed data and easy-to-follow implementation details.

In the pursuit of enhancing smart city infrastructure, computer vision serves as a pivotal element for traffic management, scene understanding, and security applications. This research investigates the performance of the YOLO v9-c and YOLO v9-e object detection models in identifying vehicles under snowy weather conditions, leveraging various data augmentation techniques. The study highlights that, historically, object detection relied on complex, handcrafted features, but deep learning advancements have enabled more efficient and accurate end-to-end learning directly from raw data. Despite these advancements, detecting objects in adverse weather conditions like snow remains challenging, affecting the safety and effectiveness of autonomous systems. The study examines the performance of YOLO v9-c and YOLO v9-e under four different scenarios: no augmentation, snow accumulation, snow overlay, and snow depth mapping. Results indicate that both models achieve their highest precision without augmentation, with YOLO v9-c and YOLO v9-e reaching precisions of 82% and 80%, respectively. However, the snow accumulation method severely impacts detection accuracy, with precision dropping to 36% for YOLO v9-c and 43% for YOLO v9-e. Snow overlay augmentation shows better adaptability, with YOLO v9-c achieving 68% and YOLO v9-e 76% precision. Snow depth mapping results in moderate impacts, with precisions of 59% for YOLO v9-c and 61% for YOLO v9-e. The findings emphasize the importance of careful selection and tuning of augmentation techniques to improve object detection models’ robustness under snowy weather conditions, thereby enhancing the safety and efficiency of autonomous systems. The study suggests a tunned augmentation that helps YOLO v9-c and YOLO v9-e reach precisions of 85% and 83%. Future research should focus more on optimizing augmentation parameters, diversifying training data, and employing domain randomization to further enhance the robustness and generalization capabilities of these models. This approach aims to ensure more reliable performance of autonomous systems in real-world conditions where adverse weather is a common occurrence. The code and the dataset will be available at https://nvd.Itu-ai.dev/

Intelligent surveillance systems are inherently computationally intensive. And with their ever-expanding utilization in both small-scale home security applications and on the national scale, the necessity for efficient computer vision processing is critical. To this end, we propose a framework that utilizes modern hardware by incorporating multi-threading and concurrency to facilitate the complex processes associated with object detection, tracking, and identification, enabling lower-powered systems to support such intelligent surveillance systems effectively. The proposed architecture provides an adaptable and robust processing pipeline, leveraging the thread pool design pattern. The developed method can achieve respectable throughput rates on low-powered or constrained compute platforms.

For an intelligent transportation system, identifying license plate numbers in drone photos is difficult, and it is used in practical applications like parking management, traffic management, automatically organizing parking spots, etc. The primary goal of the work that is being presented is to demonstrate how to extract robust and invariant features from PCM that can withstand the difficulties posed by drone images. After that, the work will take advantage of a fully connected neural network to tackle the difficulties of fixing precise bounding boxes regardless of orientations, shapes, and text sizes. The proposed work will be able to find the detected text for both license plate numbers and natural scene images which will lead to a better recognition stage. Both our drone dataset (Mimos) and the benchmark license plate dataset (Medialab) are used to assess the effectiveness of the study that has been done. To show that the suggested system can detect text of natural scenes in a wide variety of situations. Four benchmark datasets, namely, SVT, MSRA-TD-500, ICDAR 2017 MLT, and Total Text are used for the experimental results. We also describe trials that demonstrate robustness to varying height distances and angles. This work's code and data will be made publicly available on GitHub.

Targeting the current Covid 19 pandemic situation, this paper identifies the need of crowd management. Thus, it proposes an effective and efficient real-time human detection and counting solution specifically for shopping malls by producing a system with graphical user interface and management functionalities. Besides, it comprehensively reviews and compares the existing techniques and similar systems to select the ideal solution for this scenario. Specifically, advanced deep learning computer vision techniques are decided by using YOLOv3 for detecting and classifying the human objects with DeepSORT tracking algorithm to track each detected human object and perform counting using intrusion line judgment. Additionally, it converts the pretrained YOLOv3 into TensorFlow format for better and faster real-time computation using graphical processing unit instead of using central processing unit as the traditional target machine. The experimental results have proven this implementation combination to be 91.07% accurate and real-time capable with testing videos from the internet to simulate the shopping mall entrance scenario.

License plate number detection in drone images is a complex problem because the images are generally captured at oblique angles and pose several challenges like perspective distortion, non-uniform illumination effect, degradations, blur, occlusion, loss of visibility etc. Unlike, most existing methods that focus on images captured by orthogonal direction (head-on), the proposed work focuses on drone text images. Inspired by the Phase Congruency Model (PCM), which is invariant to non-uniform illuminations, contrast variations, geometric transformation and to some extent to distortion, we explore the combination of DCT and PCM (DCT-PCM) for detecting license plate number text in drone images. Motivated by the strong discriminative power of deep learning models, the proposed method exploits fully connected neural networks for eliminating false positives to achieve better detection results. Furthermore, the proposed work constructs working model that fits for real environment. To evaluate the proposed method, we use our own dataset captured by drones and benchmark license plate datasets, namely, Medialab for experimentation. We also demonstrate the effectiveness of the proposed method on benchmark natural scene text detection datasets, namely, SVT, MSRA-TD-500, ICDAR 2017 MLT and Total-Text.

This paper proposes a simple yet effective method for anomaly detection in natural scene images improving natural scene text detection and recognition. In the last decade, there has been significant progress towards text detection and recognition in natural scene images. However, in cases where there are logos, company symbols, or other decorative elements for text, existing methods do not perform well. This work considers such misclassified components, which are part of the text as anomalies, and presents a new idea for detecting such anomalies in the text for improving text detection and recognition in natural scene images. The proposed method considers the result of the existing text detection method as input for segmenting characters or components based on saliency map and rough set theory. For each segmented component, the proposed method extracts feature from the saliency map based on density, pixel distribution, and phase congruency to classify text and non-text components by exploring a fuzzy-based classifier. To verify the effectiveness of the method, we have performed experiments on several benchmark datasets of natural scene text detection, namely, MSRATD-500 and SVT. Experimental results show the efficacy of the proposed method over the existing ones for text detection and recognition in these datasets.

This paper addresses the critical need for advanced real-time vehicle detection methodologies in Vehicle Intelligence Systems (VIS), especially in the context of using Unmanned Aerial Vehicles (UAVs) for data acquisition in severe weather conditions, such as heavy snowfall typical of the Nordic region. Traditional vehicle detection techniques, which often rely on custom-engineered features and deterministic algorithms, fall short in adapting to diverse environmental challenges, leading to a demand for more precise and sophisticated methods. The limitations of current architectures, particularly when deployed in real-time on edge devices with restricted computational capabilities, are highlighted as significant hurdles in the development of efficient vehicle detection systems. To bridge this gap, our research focuses on the formulation of an innovative approach that combines the fractional B-spline wavelet transform with a tailored U-Net architecture, operational on a Raspberry Pi 4. This method aims to enhance vehicle detection and localization by leveraging the unique attributes of the NVD dataset, which comprises drone-captured imagery under the harsh winter conditions of northern Sweden. The dataset, featuring 8450 annotated frames with 26,313 vehicles, serves as the foundation for evaluating the proposed technique. The comparative analysis of the proposed method against state-of-the-art detectors, such as YOLO and Faster RCNN, in both accuracy and efficiency on constrained devices, emphasizes the capability of our method to balance the trade-off between speed and accuracy, thereby broadening its utility across various domains.

Text-to-Image synthesis is the task of generating an image according to a specific text description. Generative Adversarial Networks have been considered the standard method for image synthesis virtually since their introduction. Denoising Diffusion Probabilistic Models are recently setting a new baseline, with remarkable results in Text-to-Image synthesis, among other fields. Aside its usefulness per se, it can also be particularly relevant as a tool for data augmentation to aid training models for other document image processing tasks. In this work, we present a latent diffusion-based method for styled text-to-text-content-image generation on word-level. Our proposed method is able to generate realistic word image samples from different writer styles, by using class index styles and text content prompts without the need of adversarial training, writer recognition, or text recognition. We gauge system performance with the Fréchet Inception Distance, writer recognition accuracy, and writer retrieval. We show that the proposed model produces samples that are aesthetically pleasing, help boosting text recognition performance, and get similar writer retrieval score as real data.

This paper presents a systematic literature review of image datasets for document image analysis, focusing on historical documents, such as handwritten manuscripts and early prints. Finding appropriate datasets for historical document analysis is a crucial prerequisite to facilitate research using different machine learning algorithms. However, because of the very large variety of the actual data (e.g., scripts, tasks, dates, support systems, and amount of deterioration), the different formats for data and label representation, and the different evaluation processes and benchmarks, finding appropriate datasets is a difficult task. This work fills this gap, presenting a meta-study on existing datasets. After a systematic selection process (according to PRISMA guidelines), we select 65 studies that are chosen based on different factors, such as the year of publication, number of methods implemented in the article, reliability of the chosen algorithms, dataset size, and journal outlet. We summarize each study by assigning it to one of three pre-defined tasks: document classification, layout structure, or content analysis. We present the statistics, document type, language, tasks, input visual aspects, and ground truth information for every dataset. In addition, we provide the benchmark tasks and results from these papers or recent competitions. We further discuss gaps and challenges in this domain. We advocate for providing conversion tools to common formats (e.g., COCO format for computer vision tasks) and always providing a set of evaluation metrics, instead of just one, to make results comparable across studies.

Sign gesture recognition is the field that models sign gestures in order to facilitate communication with hearing and speech impaired people. Sign gestures are recorded with devices like a video camera or a depth camera. Palm gestures are also recorded with the Leap motion sensor. In this paper, we address palm sign gesture recognition using the Leap motion sensor. We extract geometric features from Leap motion recordings. Next, we encode the Genetic Algorithm (GA) for feature selection. Genetically selected features are fed to different classifiers for gesture recognition. Here we have used Support Vector Machine (SVM), Random Forest (RF), and Naive Bayes (NB) classifiers to have their comparative results. The gesture recognition accuracy of 74.00% is recorded with RF classifier on the Leap motion sign gesture dataset.

The use of drones for several real-world applications is increasing exponentially, especially for the purpose of monitoring, surveillance, security, etc. Most existing scene text detection methods were developed for normal scene images. This work aims to develop a model for detecting text in drone as well as scene images. To reduce the adverse effects of drone images, we explore the combination of Fourier transform and Convolutional Block Attention Module (CBAM) to enhance the degraded information in the images without affecting high-contrast images. This is because the above combination helps us to extract prominent features which represent text irrespective of degradations. Therefore, the refined features extracted from the Fourier Contouring Network (FCN) are supplied to Vision Transformer, which uses the ResNet50 as a backbone and encoder-decoder for text detection in both drone and scene images. Hence, the model is called Fourier Transform based Transformer. Experimental results on drone datasets and benchmark datasets, namely, Total-Text and ICDAR 2015 of natural scene text detection show the proposed model is effective and outperforms the state-of-the-art models.

Researchers have been using Electroencephalography (EEG) to build Brain-Computer Interfaces (BCIs) systems. They have had a lot of success modeling brain signals for applications, including emotion detection, user identification, authentication, and control. The goal of this study is to employ EEG-based neurological brain signals to recognize imagined objects. The user imagines the object after looking at the same on the monitor screen. The EEG signal is recorded when the user thinks up about the object. These EEG signals were processed using signal processing methods, and machine learning algorithms were trained to classify the EEG signals. The study involves coarse and fine level EEG signal classification. The coarse-level classification categorizes the signals into three classes (Char, Digit, Object), whereas the fine-level classification categorizes the EEG signals into 30 classes. The recognition rates of 97.30%, and 93.64% were recorded at coarse and fine level classification, respectively. Experiments indicate the proposed work outperforms the previous methods.

The rising diversity, volume, and pace of fashion manufacturing pose a considerable challenge in the fashion industry, making it difficult for customers to pick which product to purchase. In addition, fashion is an inherently subjective, cultural notion and an ensemble of clothing items that maintains a coherent style. In most of the domains in which Recommender Systems are developed (e.g., movies, e-commerce, etc.), the similarity evaluation is considered for recommendation. Instead, in the Fashion domain, compatibility is a critical factor. In addition, raw visual features belonging to product representations that contribute to most of the algorithm’s performances in the Fashion domain are distinguishable from the metadata of the products in other domains. This literature review summarizes various Artificial Intelligence (AI) techniques that have lately been used in recommender systems for the fashion industry. AI enables higher-quality recommendations than earlier approaches. This has ushered in a new age for recommender systems, allowing for deeper insights into user-item relationships and representations and the discovery patterns in demographical, textual, virtual, and contextual data. This work seeks to give a deeper understanding of the fashion recommender system domain by performing a comprehensive literature study of research on this topic in the past 10 years, focusing on image-based fashion recommender systems taking AI improvements into account. The nuanced conceptions of this domain and their relevance have been developed to justify fashion domain-specific characteristics.

Debates persist regarding the impact of Stain Normalization (SN) on recent breast cancer histopathological studies. While some studies propose no influence on classification outcomes, others argue for improvement. This study aims to assess the efficacy of SN in breast cancer histopathological classification, specifically focusing on Invasive Ductal Carcinoma (IDC) grading using Convolutional Neural Networks (CNNs). The null hypothesis asserts that SN has no effect on the accuracy of CNN-based IDC grading, while the alternative hypothesis suggests the contrary. We evaluated six SN techniques, with five templates selected as target images for the conventional SN techniques. We also utilized seven ImageNet pre-trained CNNs for IDC grading. The performance of models trained with and without SN was compared to discern the influence of SN on classification outcomes. The analysis unveiled a p-value of 0.11, indicating no statistically significant difference in Balanced Accuracy Scores between models trained with StainGAN-normalized images, achieving a score of 0.9196 (the best-performing SN technique), and models trained with non-normalized images, which scored 0.9308. As a result, we did not reject the null hypothesis, indicating that we found no evidence to support a significant discrepancy in effectiveness between stain-normalized and non-normalized datasets for IDC grading tasks. This study demonstrates that SN has a limited impact on IDC grading, challenging the assumption of performance enhancement through SN.

Computer-aided Invasive Ductal Carcinoma (IDC) grading classification systems based on deep learning have shown that deep learning may achieve reliable accuracy in IDC grade classification using histopathology images. However, there is a dearth of comprehensive performance comparisons of Convolutional Neural Network (CNN) designs on IDC in the literature. As such, we would like to conduct a comparison analysis of the performance of seven selected CNN models: EfficientNetB0, EfficientNetV2B0, EfficientNetV2B0-21k, ResNetV1-50, ResNetV2-50, MobileNetV1, and MobileNetV2 with transfer learning. To implement each pre-trained CNN architecture, we deployed the corresponded feature vector available from the TensorFlowHub, integrating it with dropout and dense layers to form a complete CNN model. Our findings indicated that the EfficientNetV2B0-21k (0.72B Floating-Point Operations and 7.1 M parameters) outperformed other CNN models in the IDC grading task. Nevertheless, we discovered that practically all selected CNN models perform well in the IDC grading task, with an average balanced accuracy of 0.936 ± 0.0189 on the cross-validation set and 0.9308 ± 0.0211on the test set.

Despite the recent progress on scaling multilingual machine translation (MT) to severalunder-resourced African languages, accuratelymeasuring this progress remains challenging,since evaluation is often performed on n-grammatching metrics such as BLEU, which typically show a weaker correlation with humanjudgments. Learned metrics such as COMEThave higher correlation; however, the lack ofevaluation data with human ratings for underresourced languages, complexity of annotationguidelines like Multidimensional Quality Metrics (MQM), and limited language coverageof multilingual encoders have hampered theirapplicability to African languages. In this paper, we address these challenges by creatinghigh-quality human evaluation data with simplified MQM guidelines for error detection and direct assessment (DA) scoring for 13 typologically diverse African languages. Furthermore, we develop AFRICOMET: COMETevaluation metrics for African languages byleveraging DA data from well-resourced languages and an African-centric multilingual encoder (AfroXLM-R) to create the state-of-theart MT evaluation metrics for African languages with respect to Spearman-rank correlation with human judgments (0.441).

This paper investigates the application of large language models (LLMs) in the educational field, specifically focusing on roles like "Teacher Assistant" and "Study Assistant" to enhance personalized and adaptive learning. The significance of integrating AI in educational frameworks is underscored, given the shift towards AI-powered educational tools. The methodology of this research is structured and multifaceted, examining the dynamics between prompt engineering, methodological approaches, and LLM outputs with the help of indexed documents. The study bifurcates its approach into prompt structuring and advanced prompt engineering techniques. Initial investigations revolve around persona and template prompts to evaluate their individual and collective effects on LLM outputs. Advanced techniques, including few-shot and chain-of-thought prompting, are analyzed for their potential to elevate the quality and specificity of LLM responses. The "Study Assistant" aspect of the study involves applying these techniques to educational content across disciplines such as biology, mathematics, and physics. Findings from this research are poised to contribute significantly to the evolution of AI in education, offering insights into the variables that enhance LLM performance. This paper not only enriches the academic discourse on LLMs but also provides actionable insights for the development of sophisticated AI-based educational tools. As the educational landscape continues to evolve, this research underscores the imperative for continuous exploration and refinement in the application of AI to fully realize its benefits in education.