In this work, deep neural networks made up of multiple hidden Long Short-Term Memory (LSTM) and Feedforward layers are trained to predict the thermal behavior of the joint motors of robot manipulators. A model-free and scalable approach is adopted. It accommodates complexity and uncertainty challenges stemming from the derivation, identification, and validation of a large number of parameters of an approximation model that is hardly available. To this end, sensed joint torques are collected and processed to foresee the thermal behavior of joint motors. Promising prediction results of the machine learning based capture of the temperature dynamics of joint motors of a redundant robot with seven joints are presented.

Abstract-- Dementia is a progressive condition that impairs an individual's cognitive health and daily functioning, with mild cognitive impairment (MCI) often serving as its precursor. The prediction of MCI-to-dementia conversion has been well studied, but previous studies have almost always focused on traditional Machine Learning (ML)--based methods that require sharing sensitive clinical information to train predictive models. This work highlights that FL can eliminate the need for data sharing without compromising model efficacy. An early dementia diagnosis is essential for guiding appropriate management strategies and implementing timely I. Predicting loss of the structure and functions of the neurons, resulting whether an individual suffering from MCI will have a in a diverse group of disorders such as Alzheimer's disease, dementia diagnosis in future has been considered to be a Parkinson's disease and others. Neurodegenerative diseases key aspect towards early dementia diagnosis and large-scale cause a decrease in cognitive functions, affecting memory studies on this MCI-to-dementia conversion prediction are and/or behavioral abilities, finally interfering with the quality clearly warranted.

MLPs have been one of key components in modern neural network architectures for years. Their simplicity makes them widely used for capturing complex relationships through multiple layers of non-linear transformations. However, their role has been reconsidered recently with the revival of Kolmogorov-Arnold Networks (KANs) [1, 2]. In these papers, fixed activation functions used in MLPs are described as "nodes," and the authors proposed replacing them with learnable activation functions like B-splines, referred to as "edges", to improve performance in mathematical and physical examples. To address Hilbert's 13th problem [3], the Kolmogorov-Arnold Representation Theorem (KART) [4] was introduced. It posits that any continuous function involving multiple variables can be decomposed into a sum of continuous functions of single variables, thus inspiring the creation of KANs. The work of Liu et al. [1] on KANs has inspired numerous studies exploring the use of various basis and polynomial functions as replacements for B-splines [5, 6, 7, 8, 9, 10, 11, 12, 13], investigating the model's performance compared to MLPs. Several studies have shown that KANs do not always outperform MLPs when using the same training parameters [14, 15]. Moreover, while KANs achieve better performance than MLPs with the same network structure, they often require a significantly larger number of parameters [7, 16, 17, 18, 19].

Data pre-processing is a significant step in machine learning to improve the performance of the model and decreases the running time. This might include dealing with missing values, outliers detection and removing, data augmentation, dimensionality reduction, data normalization and handling the impact of confounding variables. Although it is found the steps improve the accuracy of the model, but they might hinder the explainability of the model if they are not carefully considered especially in medicine. They might block new findings when missing values and outliers removal are implemented inappropriately. In addition, they might make the model unfair against all the groups in the model when making the decision. Moreover, they turn the features into unitless and clinically meaningless and consequently not explainable. This paper discusses the common steps of the data preprocessing in machine learning and their impacts on the explainability and interpretability of the model. Finally, the paper discusses some possible solutions that improve the performance of the model while not decreasing its explainability.

Adapters provide an efficient and lightweight mechanism for adapting trained transformer models to a variety of different tasks. However, they have often been found to be outperformed by other adaptation mechanisms, including low-rank adaptation. In this paper, we provide an in-depth study of adapters, their internal structure, as well as various implementation choices. We uncover pitfalls for using adapters and suggest a concrete, improved adapter architecture, called Adapter+, that not only outperforms previous adapter implementations but surpasses a number of other, more complex adaptation mechanisms in several challenging settings. Despite this, our suggested adapter is highly robust and, unlike previous work, requires little to no manual intervention when addressing a novel scenario. Adapter+ reaches state-of-the-art average accuracy on the VTAB benchmark, even without a per-task hyperparameter optimization.

Despite its exceptional soft tissue contrast, Magnetic Resonance Imaging (MRI) faces the challenge of long scanning times compared to other modalities like X-ray radiography. Shortening scanning times is crucial in clinical settings, as it increases patient comfort, decreases examination costs and improves throughput. Recent advances in compressed sensing (CS) and deep learning allow accelerated MRI acquisition by reconstructing high-quality images from undersampled data. While reconstruction algorithms have received most of the focus, designing acquisition trajectories to optimize reconstruction quality remains an open question. This thesis explores two approaches to address this gap in the context of Cartesian MRI. First, we propose two algorithms, lazy LBCS and stochastic LBCS, that significantly improve upon G\"ozc\"u et al.'s greedy learning-based CS (LBCS) approach. These algorithms scale to large, clinically relevant scenarios like multi-coil 3D MR and dynamic MRI, previously inaccessible to LBCS. Additionally, we demonstrate that generative adversarial networks (GANs) can serve as a natural criterion for adaptive sampling by leveraging variance in the measurement domain to guide acquisition. Second, we delve into the underlying structures or assumptions that enable mask design algorithms to perform well in practice. Our experiments reveal that state-of-the-art deep reinforcement learning (RL) approaches, while capable of adaptation and long-horizon planning, offer only marginal improvements over stochastic LBCS, which is neither adaptive nor does long-term planning. Altogether, our findings suggest that stochastic LBCS and similar methods represent promising alternatives to deep RL. They shine in particular by their scalability and computational efficiency and could be key in the deployment of optimized acquisition trajectories in Cartesian MRI.

Data normalization is an essential task when modeling a classification system. When dealing with data streams, data normalization becomes especially challenging since we may not know in advance the properties of the features, such as their minimum/maximum values, and these properties may change over time. We compare the accuracies generated by eight well-known distance functions in data streams without normalization, normalized considering the statistics of the first batch of data received, and considering the previous batch received. We argue that experimental protocols for streams that consider the full stream as normalized are unrealistic and can lead to biased and poor results. Our results indicate that using the original data stream without applying normalization, and the Canberra distance, can be a good combination when no information about the data stream is known beforehand.

Federated learning (FL) facilitates collaborative learning among multiple clients in a distributed manner, while ensuring privacy protection. However, its performance is inevitably degraded as suffering data heterogeneity, i.e., non-IID data. In this paper, we focus on the feature distribution skewed FL scenario, which is widespread in real-world applications. The main challenge lies in the feature shift caused by the different underlying distributions of local datasets. While the previous attempts achieved progress, few studies pay attention to the data itself, the root of this issue. Therefore, the primary goal of this paper is to develop a general data augmentation technique at the input level, to mitigate the feature shift. To achieve this goal, we propose FedRDN, a simple yet remarkably effective data augmentation method for feature distribution skewed FL, which randomly injects the statistics of the dataset from the entire federation into the client's data. By this, our method can effectively improve the generalization of features, thereby mitigating the feature shift. Moreover, FedRDN is a plug-and-play component, which can be seamlessly integrated into the data augmentation flow with only a few lines of code. Extensive experiments on several datasets show that the performance of various representative FL works can be further improved by combining them with FedRDN, which demonstrates the strong scalability and generalizability of FedRDN. The source code will be released.

With high device integration density and evolving sophisticated device structures in semiconductor chips, detecting defects becomes elusive and complex. Conventionally, machine learning (ML)-guided failure analysis is performed with offline batch mode training. However, the occurrence of new types of failures or changes in the data distribution demands retraining the model. During the manufacturing process, detecting defects in a single-pass online fashion is more challenging and favoured. This paper focuses on novel quantile online learning for semiconductor failure analysis. The proposed method is applied to semiconductor device-level defects: FinFET bridge defect, GAA-FET bridge defect, GAA-FET dislocation defect, and a public database: SECOM. From the obtained results, we observed that the proposed method is able to perform better than the existing methods. Our proposed method achieved an overall accuracy of 86.66% and compared with the second-best existing method it improves 15.50% on the GAA-FET dislocation defect dataset.

Data will always give us the best results if we treat it in the right way that's why it has a better idea than dealing with it directly. In this article, we will discuss feature scaling, The most commonly feature scaling methods and when to use them. Feature scaling is a significant stage in the pre-processing of data before developing a machine learning model. The difference between a bad and a good machine learning model can be determined by scaling. Machine learning algorithms deals with numbers only, and if there is a significant difference in range, such as a few ranging in the hundreds against a few ranging in the tens, it assumes that greater ranging numbers have some form of superiority.