Our study develops a generative adversarial network (GAN)-based method that generates faithful synthetic image data of human cardiomyocytes at varying stages in their maturation process, as a tool to significantly enhance the classification accuracy of cells and ultimately assist the throughput of computational analysis of cellular structure and functions. Methods: Human induced pluripotent stem cell derived cardiomyocytes (hiPSC-CMs) were cultured on micropatterned collagen coated hydrogels of physiological stiffnesses to facilitate maturation and optical measurements were performed for their structural and functional analyses. Control groups were cultured on collagen coated glass well plates. These image recordings were used as the real data to train the GAN model. Results: The results show the GAN approach is able to replicate true features from the real data, and inclusion of such synthetic data significantly improves the classification accuracy compared to usage of only real experimental data that is often limited in scale and diversity. Conclusion: The proposed model outperformed four conventional machine learning algorithms with respect to improved data generalization ability and data classification accuracy by incorporating synthetic data. Significance: This work demonstrates the importance of integrating synthetic data in situations where there are limited sample sizes and thus, effectively addresses the challenges imposed by data availability.

In the realm of virtual reality (VR), immersive experiences are enhanced by sophis=cated tracking and inference technologies that me=culously monitor user movements, interac=ons, and behaviors. While these advancements offer unprecedented levels of engagement and personaliza=on, they also raise significant privacy concerns. This study explores the dual-edged nature of tracking and inference in VR, examining how these technologies influence user percep=ons of privacy, trust, and comfort within virtual environments. Through a mixed-methods approach that combines quan=ta=ve surveys with qualita=ve interviews, this research delves into the aJtudes of VR users towards the data collec=on prac=ces embedded within these digital realms. The study inves=gates the awareness levels among users regarding the extent of tracking and inference, their reac=ons to such prac=ces, and the poten=al impact on their con=nued use of VR technologies.

ECG-imaging has been receiving increasing clinical and commercialization interest as a non-invasive technology for computing cardiac electrical activity and facilitating clinical management of heart rhythm disorders. Despite its potential, the standard ECG-imaging pipeline requires thorax imaging for constructing patient-specific heart-thorax geometry – being outside standard-of-care clinical workflow, this component constitutes a major barrier to the clinical adoption of ECG-imaging. The advent of 3D cameras into ECG-imaging workflow to replace thorax imaging has shown promise, although existing works largely neglect the registration between camera-derived thorax models with an individual’s heart geometry. In this work, we address this gap with a novel system that generates patient-specific torso geometry using a 3D camera, registers the torso to heart geometry obtained from preexisting cardiac scans, and optimally deforms the torso to the skin surface of the patient. We evaluated the presented camera-based ECG-imaging system on five patients undergoing ablation of scar-related ventricular tachycardia, where we evaluate both the accuracy of the surface electrode localization and the ECG-imaging solutions in comparison to those obtained from computed tomography based thorax imaging. We further evaluate the use of the presented camera-based patient-specific heart-thorax model versus a generic heart-thorax model, highlighting the importance of the registration between the camera-based thorax models with cardiac scans. 

Depression is a severe psychosocial pathology that causes mood changes, characterized by a strong feeling of hopelessness and deep sadness. In advanced stages, it can predispose patients to suicidal thoughts, highlighting the importance of finding methods that provide more accurate diagnoses. Traditional diagnosis relies on semi-structured interviews and complementary questionnaires. Combining these methods with careful data analysis that incorporates audiovisual and textual characteristics can obtain valuable clues about the presence of depression in individuals. Therefore, this study proposes a multimodal Ensemble Stacking Deep Neural Network model based on the analysis of facial expression characteristics, audio signals, and textual transcriptions to automatically detect depression. A comprehensive model was evaluated on the multimodal Distress Analysis Interview Corpus-Wizard of Oz dataset. We incorporated substantial volumes of data into the analysis and achieved a degree of separability greater than 0.9. Our results demonstrate both the effectiveness of the method and its superiority to other reference approaches.

This study conducted a thorough examination of residential property data in the Abbotsford area of M elbourne with the goal of identifying significant trends in housing, regional patterns, and variables aff ecting home values.The dataset contained a number of elements, such as neighborhood information, g eographic coordinates, transaction details, and property attributes.Using a variety of techniques, inclu ding mean imputation, forward filled imputation, machine learning algorithms, and discarding missin g data, the study started with the identification and treatment of missing values.Particularly in the pric e variable, outliers were found, and boxplots and other visualization tools were used for outlier analysi s.For additional analysis, numerical values representing the categorical variables were converted.To in vestigate the distributions of numerical variables and comprehend connections between variables with a focus on correlations with home prices-univariate and bivariate analyses were carried out. Feature engineering, covariance analysis, ANOVA testing, and predictive modeling with regression algorithms like Random Forest, XGBoost, and Support Vector Machine (SVM) were all part of the quantitative analysis process. Metrics like Mean Absolute Error (MAE) were used to assess the performance of the model; the results showed that XGBoost was the most accurate predictor of housing prices. Significant factors influencing home prices were identified by the study, such as building area, property type, number of rooms, and geographic considerations including proximity to important sites. Each component was analyzed in terms of its relative relevance, and the building area and land size. It was noted how the constructed model has limits, such as overfitting and the need for more model refining. The results offer insightful information to scholars, politicians, and real estate professionals who are interested in the dynamics of the housing market.

In electrical impedance tomography (EIT) systems, various current sources were designed to deliver constant electrical current towards the electrodes, irrespective of the conductivity of the conductive medium under test. This, however, is not possible to achieve in practice since the water conductivity ranges from 5 S/m for sea water down-to 5.5 x10-6 and 5 x 10-3 S/m for pure and drinking water respectively which are substantially low values [1]. Thus, even in case of high water-cut fluid, unless the water is very salty, the usage of such current sources may not be appropriate and their substitution with an electrical capacitance tomography system or a dielectric measurement sensor may be required. Indeed, even if the conductivity of the medium is known and high, the gross conductivity formed between a given pair of electrodes depend heavily on the phase’s distribution pattern between them and can be excessively low, causing high electric current fluctuations. One of the contributions of this paper is to experimentally assess the effect of the electrical current fluctuations on the accuracy of the EIT image reconstruction using three different cutting-edge and most widely current sources designs. To the authors’ best knowledge, this study is the first of its kind as all other prior works assume that the electrical current is constant in the formulation of the EIT forward and inverse problems. The paper also suggests a new cost-effective measurement circuit design that overcomes the fluctuations. It continuously measures the electrical current consumed during every single excitation cycle using a very high-speed analog-to-digital (ADC) converter, interfaced with Cyclone V field program gate array (FPGA), to compensate for the associate voltage readings accordingly during the image reconstruction procedure. The assessment of the system was conducted experimentally, and the results were compared with those provided by the three current sources. The associated results show the higher accuracy of the suggested design, when using Gauss Newton (GN) method, in terms of mean-squared error, which was decreased by 75%, and the image correlation coefficient as well.

This study presents a comprehensive evaluation of the cybersecurity robustness of five leading Large Language Models (LLMs)-ChatGPT-4, Google Gemini, Anthropic Claude, Meta Llama, and Mistral 8x7B-against adversarial prompts using the PromptBench benchmark. Through a dual approach of quantitative and qualitative analysis, the research explores each model's performance, resilience, and vulnerabilities. Quantitative metrics such as accuracy, precision, recall, and F1 scores offer a statistical comparison across models, while qualitative insights reveal distinct patterns of response and susceptibility to various adversarial strategies. The findings highlight significant variations in model robustness, underlining the importance of a complex approach to enhancing LLM security. This study not only sheds light on current limitations but also emphasizes the need for advancing evaluation methodologies and model development practices to mitigate potential threats and ensure the safe deployment of LLMs in sensitive and critical applications.

The sums of prime components of composite numbers lead to prime numbers, a remarkable phenomena in number theory that is examined in this paper. We present a constructive proof methodology to show that adding the prime components of the resulting composite numbers iteratively converges to prime numbers for any composite number higher than four. By applying logical reasoning and the fundamentals of number theory, we prove that all prime numbers larger than three have a finite number of "prime friends," or composite numbers whose prime factors add up to the prime number itself. We reveal the method by which prime numbers emerge from this iterative process by analyzing partitions of prime numbers and applying the function 2k + 3. This abstract summarizes our investigation into the convergence of prime numbers through the sums of prime factors of composite numbers, demonstrating a significant relationship between primes and composites in number theory. NOTE : We have created the term "prime friend"; it is not a term found in the traditional lexicon of mathematics.

Scene boundary detection is crucial in applications like scene segmentation and video skimming, but the presence of ad clips makes it difficult. There are some existing methods based on visual, audiovisual and audio-only features. The existing audio-only feature-based method depends on short silences between program and ad or between ads. Still, short silences can also be present inside the program, affecting performance. So we proposed a detection method for ad shots using MFCC features. MFCC is a basic speech feature based on short-term spectral. We used the average of MFCC for a shot and determined the threshold for detection purposes. We used four episodes of The Big Bang Theory as our dataset. Our method is the first method that uses MFCC features for ad detection and explores the relation between video contents and MFCC features.

Electromagnetic (EM) body models predict the impact of human presence and motions on the Radio-Frequency (RF) field originated from wireless devices nearby. Despite their accuracy, EM models are time-consuming methods which prevent their adoption in strict real-time computational imaging problems and Bayesian estimation, such as passive localization, RF tomography, and holography. Physics-informed Generative Neural Network (GNN) models have recently attracted a lot of attention thanks to their potential to reproduce a process by incorporating relevant physical laws and constraints. They can be used to simulate or reconstruct missing data or samples, reproducing EM propagation effects, approximated EM fields and learn a physics-informed data distribution, i.e., a Bayesian prior. The paper discusses two popular techniques, namely variational auto-encoders (VAEs) and generative adversarial networks (GANs), and their adaptations to incorporate relevant EM body diffraction concepts. Proposed EM-informed generative models are verified against classical EM tools driven by diffraction theory and validated on real data. Physics-informed generative machine learning represents a multidisciplinary research area weaving together physical/EM modelling, signal processing and artificial intelligence (AI): the paper explores emerging opportunities of GNN tools targeting real-time passive RF sensing in multipleinput multiple-output (MIMO) communication systems. Proposed generative tools are designed, implemented and verified on resource constrained wireless devices being members of a Wireless Local Area Network (WLAN).

In this paper, we investigate the performance of a Hybrid Quantum Neural Network (HQNN) and a comparable classical Convolution Neural Network (CNN) for detection and classification problem using a radar. Specifically, we take a fairly complex radar time-series model derived from electromagnetic theory, namely the Martin-Mulgrew model, that is used to simulate radar returns of objects with rotating blades, such as drones. We find that when that signal-to-noise ratio (SNR) is high, CNN outperforms the HQNN for detection and classification. However, in the low SNR regime (which is of greatest interest in practice) the performance of HQNN is found to be superior to that of the CNN of a similar architecture.

Large Language Models (LLMs) encounter challenges with the unique syntax of specific domains, such as biomolecules. Existing fine-tuning or modality alignment techniques struggle to bridge the domain knowledge gap and understand complex molecular data, limiting LLMs' progress in specialized fields. To overcome these limitations, we propose an expandable and adaptable nonparametric knowledge injection framework named Domain-specific Retrieval-Augmented Knowledge (DRAK), aimed at enhancing reasoning capabilities in specific domains. Utilizing knowledge-aware prompts and gold labelinduced reasoning, DRAK has developed profound expertise in the molecular domain and the capability to handle a broad spectrum of analysis tasks. We evaluated two distinct forms of DRAK variants, proving that DRAK exceeds previous benchmarks on six molecular tasks within the Mol-Instructions dataset. Extensive experiments have underscored DRAK's formidable performance and its potential to unlock molecular insights, offering a unified paradigm for LLMs to tackle knowledge-intensive tasks in specific domains. Our code will be available soon.

Quantitative assessment of spinal motion plays a pivotal role in diagnosing and understanding lower back pain. This paper utilises a Convolutional Neural Network for precise landmark localisation of bounding boxes encompassing the lumbar spine in sagittal plane lumbar fluoroscopy image sequences. The proposed methodology aims to automate spinal movement tracking and provide a benchmark for future research, thereby enhancing the efficiency and accuracy of low back pain diagnosis.

In an era of significant economic volatility, time series forecasting is widely used to predict stock prices and guide investors in trading decisions. Nevertheless, existing data-driven techniques are unable to effectively handle the vast amount of financial data due to big data constraints such as nonlinearity, non-stationarity, heteroskedasticity, and unsynchronicity. A cohesive framework is also required for ensuring the smooth integration and synchronization of varied methodologies in timeseries financial prediction tasks. To address this problem, this paper introduces a novel framework that investigates three ensemble strategies: blending, stacking, and voting, and selects the best method to perform the stock trend prediction task. Specifically, we deploy four distinct machine learning algorithms as the base learning model, each of which is uncorrelated and proficient in a different way depending on the task. The outputs of the basis classifiers are then combined using the adaptive boosting algorithm, a meta classifier, to give the final prediction results. To augment predictive models's accuracy and generalization capabilities, we put forward strategies like feature engineering and Ridge regularization, which optimize the pertinence of data and curb overfitting. Our examination of five distinct case studies on Toronto Stock Exchange data reveals that the proposed multimodel ensemble method has superior performance compared to others.

This study introduces an innovative method for forecasting corn yield and price variations, critical for food security and strategic planning. Unlike traditional methods reliant on pre-harvest production data, which are often difficult to access, our approach utilizes satellite-derived gross primary production (GPP) data and dimension-reduction techniques to predict national corn yield and price changes. We conducted case studies in the US, Malawi, and South Africa to validate this approach, extracting predictors from annual GPP variations during peak growing seasons. We utilized various dimension reduction strategies, including spatial averaging, Empirical Orthogonal Functions (EOFs), and deep learning approaches like Autoencoders (AEs) and Variational Autoencoders (VAEs), to extract meaningful features from the GPP datasets. These extracted features serve as predictors in statistical models such as Generalized Linear Models and the Least Absolute Shrinkage and Selection Operator (LASSO), along with a neural network trained to predict variations directly from GPP-derived latent features. Model performances were evaluated using the Area Under Curve, Brier Skill Score, and Matthew Correlation Coefficient. Our results indicate that neural network models based on both AEs and VAEs exhibit superior predictive capabilities across all three countries, with the VAE excelling in the US and the AE leading in South Africa and Malawi. Similarly, the VAE outperforms other approaches for price prediction in the US and South Africa, while the AE achieves the best results in Malawi. Our study demonstrates the potential of combining satellite data and dimension reduction methods to significantly enhance large-scale corn yield and price predictions months before harvest.

Convolutional Neural Networks (CNNs) have opened up new possibilities in a variety of fields, including plant pathology. Through image analysis, this study aims to take advantage of CNNs' capacity to recognize and categorize plant illnesses. Automated disease identification has potential for prompt intervention and crop preservation in modern times, when agricultural output assumes critical relevance. This work elaborates on the application and assessment of a CNN-based model designed for the detection of plant diseases. The technology separates healthy plants from those with problems by taking pictures of plants and putting them through the trained model. The growing demand for effective disease monitoring and management in the agriculture industry highlights the urgency of this research. The suggested model shows promising accuracy and dependability by utilising a large dataset and stringent evaluation metrics. This research highlights the potential of CNNs as a workable tool for upgrading agricultural practices by using a systematic approach. The results of this study shed light on both the strengths and weaknesses of CNNs in the context of automated plant disease diagnosis.

Despite the impressive performance of current deep learning models in the field of medical imaging, the transfer of the lung segmentation task in X-ray images to clinical practice is still a pending task. In this study, we explored the performance of a fully automatic framework for lung fields segmentation in chest X-ray images, based on the combination of the Segment Anything Model (SAM) with prompt capabilities, and the You Only Look Once (YOLO) model to provide effective prompts. Transfer learning, loss functions and several validation strategies were intensively assessed. This provides a complete benchmark that enables future research studies to fairly compare new segmentation strategies. The results achieved demonstrate significant robustness and generalization capability against the variability in sensors, populations, disease manifestations, device processing and imaging conditions. The proposed framework is computationally efficient, can address bias in training over multiple datasets, and has the potential to be applied across other domains and modalities.

In this paper, a new method based on machine learning methods is used to find prime numbers. In this way, by training a regression model using the sum of the previous prime numbers, the next prime number can be calculated with a very small error. Even the next one hundred prime numbers can be guessed. The error of this calculation is very small and this error is alternating (sine or cosine) and it becomes zero in some places. Therefore, it can be said that due to zero error, the next prime number must exist and this theorem proves the infinity of prime numbers.

In situations where visible lighting is inadequate for sensing, infrared sensors are commonly employed. However, they often yield blurry images lacking clear textures and terrain/object boundaries. Unfortunately, human visibility diminishes even with infrared sensors providing more visual information than visible sensors, especially at night aerial imagery. To enhance the visibility of aerial infrared images, we propose adopting semantic segmentation, which assigns pixel-wise class labels to various input images, thereby clarifying substantial boundaries. However, training an accurate semantic segmentation model necessitates extensive pixel-wise annotations corresponding to input images, which are lacking in aerial infrared images with ground truth datasets. To address this challenge, we introduce a novel method that automatically generates pixel-wise class labels using solely infrared images and metadata such as GPS coordinates. Our method comprises two pivotal functions: coarse alignment with metadata in geographic information system (GIS) space and fine alignment based on multimodal image registration between aerial images. Aerial image datasets spanning three domains-day, twilight, and night-were created using shortwave infrared (SWIR) and mid-wave infrared (MWIR) images captured by optical sensors mounted on helicopters. Experimental results demonstrate that training on GIS data as label images enables high-precision semantic segmentation across both daytime and nighttime conditions.

In the field of few-shot learning, the increasing emergence of novel relations poses a significant challenge to existing models that leverage knowledge from past relation classification tasks. This issue becomes more pronounced when the volume of novel relations overtakes that of existing ones, underscoring the need for models that are less dependent on base relations and more adept at learning from the novel ones. This paper addresses the challenging Class-Incremental Fewshot Relation Classification (CIFRC) problem by a model named ReliDispProto: Reliability and Dispersion-infused Prototypical network. ReliDispProto employs a sophisticated similarity metric that accounts for the reliability of query-to-prototype matches and the distribution dispersion of support classes. Based on the metric, unlabeled instances are efficiently classified into support classes across a series of sessions. While training on novel relations, ReliDispProto is fine-tuned by classification reliability assessed in each iteration. To further enhance its performance, we integrate teacher-student knowledge distillation and label smoothing techniques. These additions effectively alleviate issues such as catastrophic forgetting and overfitting. Experimental results on three public datasets reveal that ReliDispProto significantly outperforms existing state-of-the-art methods, achieving up to 15.63% improvements in accuracy. The datasets and source code for ReliDispProto are accessible from http://github.com/nickhcx/ReliDispProto.

The paper focuses on the multi-azimuth interpolation task of inverse synthetic aperture radar (ISAR) images for aircraft targets and complements incomplete ISAR image datasets. ISAR image automatic target recognition (ATR) has been widely applied in remote sensing and many fields. However, the imaging process is more challenging when compared to capturing optical and SAR image data, which reduces the accuracy and generalization performance of the ATR system. Therefore, in this paper, we leverage existing limited ISAR data to achieve autonomous data expansion. This approach helps mitigate the impact of low sample quantity and unbalanced distribution, ultimately improving the accuracy of the ATR system for target recognition. Most existing methods use generative networks for ISAR image expansion, but few focus on generating ISAR images with specific azimuths. The paper proposes a novel two-stage coarse-to-fine framework for ISAR object view interpolation (C2FIPNet) that combines flow estimation and GAN to interpolate ISAR images with intermediate azimuths using a set of ISAR image pairs. Flow estimation is employed for coarsegrained generation, determining the position and intensity of strong scattering points in the ISAR image. The GAN, on the other hand, is used for fine-grained completion to correct image distortion caused by flow estimation and enhance image details. Additionally, a suitable loss function is designed, incorporating both global and local features, allowing for priority generation in the region of strong scattering points. In conclusion, extensive simulation and comparative experiments have demonstrated that the interpolated ISAR images generated by the proposed C2FIPNet exhibit greater pixel-level authenticity.

In this study we systematically investigate a range of nonlinear functions in the fully-connected feed-forward portions of Vision Transformer (ViT) machine learning models. We limit our investigation to only GLU-type nonlinear functions, as the GLU-type function SwiGLU is the current standard in state-ofthe-art language and vision models. We identify 21 candidate layer functions with 1, 2, or 3 weight matrices, utilizing the activations: Sigmoid, Tanh, and Sin. ViT-Tiny models are implemented utilizing these functions and benchmarked on the image classification datasets CIFAR10, CIFAR100, and SVHN. Through these experiments we identify several previously uninvestigated functions which consistently outperform SwiGLU on all benchmarks. The most performant of these functions we call SinGLU. We further benchmark SwiGLU and SinGLU on the image classification dataset Imagenet64, and again SinGLU performs better. We note that periodic functions such as Sin are not common in neural networks. We perform a numerical investigation into sinusoidally activated neurons and suggest that their viability in Vision Transformers may be due to the losslandscape smoothing of Layer Normalization, Multi-Headed Self Attention (MHSA), and modern data augmentations such as label smoothing. However, our experiments on the CIFAR100 dataset indicate that layer nonlinearity remains a hyperparameter, with approximately piecewise-linear functions performing better than more complex functions when the problem space is not densely sampled.

Data mining is an evolutionary process, and data and information are also manifestations of knowledge, but people see concepts, rules, patterns, laws and constraints more as knowledge. Based on data mining, an expert system works by reasoning using expert knowledge, and it consists of knowledge acquisition, knowledge base, reasoning machine, and interpretation part. The knowledge base is used to store experts' experience, knowledge and existing knowledge in the field, and whether it is perfect or not determines the performance of the expert system. Data mining is getting more and more attention, while the existing expert systems or decision support systems still mainly rely on manual collection and manual input of required knowledge and information, which is not only time-consuming and expensive, but also lags behind the time when making time-sensitive decisions and reduces the quality of decision-making.

This paper examines the evolving landscape of machine learning (ML) and its profound impact across various sectors, with a special focus on the emerging field of Privacy-preserving Machine Learning (PPML). As ML applications become increasingly integral to industries like telecommunications, financial technology, and surveillance, they raise significant privacy concerns, necessitating the development of PPML strategies. The paper highlights the unique challenges in safeguarding privacy within ML frameworks, which stem from the diverse capabilities of potential adversaries, including their ability to infer sensitive information from model outputs or training data. We delve into the spectrum of threat models that characterize adversarial intentions, ranging from membership and attribute inference to data reconstruction. The paper emphasizes the importance of maintaining the confidentiality and integrity of training data, outlining current research efforts that focus on refining training data to minimize privacy-sensitive information and enhancing data processing techniques to uphold privacy. Through a comprehensive analysis of privacy leakage risks and countermeasures in both centralized and collaborative learning settings, this paper aims to provide a thorough understanding of effective strategies for protecting ML training data against privacy intrusions. It explores the balance between data privacy and model utility, shedding light on privacy-preserving techniques that leverage cryptographic methods, Differential Privacy, and Trusted Execution Environments. The discussion extends to the application of these techniques in sensitive domains, underscoring the critical role of PPML in ensuring the privacy and security of ML systems.