Search

Random splitting strategy is a common approach for training, testing, and validating object detection algorithms based on deep learning. Is common for datasets to have images extracted from video sources, in which there are frames with high spatial correlation, i.e., frames with rotated positions or different view angles of the same object. These highly correlated frames may lead to information leakage in training, if these frames are not well-distributed. In this work, it is shown that datasets created with highly spatial correlation frames from the same video have information leakage if using the random splitting strategy to distribute the image into the sub-datasets. It proposed a clustering dataset split algorithm in which images are distributed randomly in the sub-datasets in a pack or clusters instead of a single image at the time. The clusters are created by extracting the image features from a video of the dataset using an image-text pre-trained model, CLIP, and reducing the feature vector dimensionality with t-Distributed Stochastic Neighbor embedding (t-SNE). In this reduced dimensional representation, images are separated into clusters using a clustering algorithms like DBSCAN, OPTICS, and Agglomerative Clustering. These clusters are distributed into the train, test, and validation datasets randomly to avoiding information leakage by highly spatial correlation frames. YOLOv8 is used as the object detector algorithm to test the dataset splitting.

The frequency of an electrical power system can be interpreted as an indicator of the equilibrium between generation and demand. Under fault conditions where this balance is disrupted, the Under-Frequency Load Shedding scheme is typically the ultimate resource employed to prevent the frequency from dropping to undesirable thresholds that could lead to a blackout. Smart loads have the capability to dynamically adjust their consumption in response to frequency variations, which has the potential to enhance frequency regulation and mitigate the need for abrupt load disconnection through the Under-Frequency Load Shedding scheme. The objective of this paper is to present a literature review of Under-Frequency Load Shedding schemes, an innovative approach to their classifications, and the introduction of a new category within these schemes that incorporates smart loads. Additionally, an overview of smart loads is included to be potentially employed in the design of new schemes and contribute to a shift in the operation philosophy, considering that the primary mission of electric power systems is to provide energy to loads rather than disconnecting them.

In tokamak and other fusion devices, magnetic control is the main tool that allows to regulate the plasma current, position and shape; it is in charge of actuating the desired plasma current waveform, steering the plasma position to a given set point and maintain the plasma shape close to a prescribed plasma equilibrium. This work describes the application of several physics concepts and computational tools in order to obtain a novel optimal controller for the plasma centroid position, which has been implemented and tested in the real-time plasma control system at the ISTTOK tokamak. A key point for the development of the new control system was the installation of a recently upgraded hardware, that numerically integrates in real-time the magnetic probes signals.

Predictive process monitoring techniques aim to forecast outcomes of running business process instances. These techniques are based on using predictive models built from past observed behavior, i.e., in an offline setting. However, business process behavior usually changes over time and predictive models are therefore at risk of becoming obsolete. Because of this, the definition of systems that build predictive models through an online setting has recently gained attention. Nevertheless, the scalability of this kind of setting within a context where the amount of data available is experiencing rapid growth is an outstanding issue. To solve this problem, this paper aims to define a framework for event sequence prediction capable of taking advantage of modern distributed processing platforms. An implementation over this framework based on Apache Flink is presented and it is tested upon two different case studies to prove its validity and its capacity to scale.

This work presents and evaluates the use of geometric parameters of the environment in the prediction of the electric field in mixed city-river type environments, employing two techniques of Machine Learning (ML) as Artificial Neural Networks (ANN) and Neuro-Fuzzy System (NFS). For its development, measurements were carried out in Amazon Region, Belém city, in the 521 MHz band. The input parameters for an ANN and NFS are the distance between transmitter and receiver, the distance only over the river, the height of the ground, the radius of the first Fresnel ellipsoid, and the electric field of free space. The ANN is a Multilayer Perceptron Network (MLP) that uses the Levenberg-Marquardt training algorithm and cross-validation method. The NFS is an Adaptive Neuro-Fuzzy Inference System (ANFIS) that uses the model Sugeno. The results obtained compared with the classic literature models (ITU-R 1546 and Okumura-Hata) in the city for distances up 20 km and over the river for distances up 5 km. A quantitative analysis is performed between the measured and predicted data through the Standard deviation (SD), Root Mean Square Error (RMSE), and the Grey Relational Grade, combined with the Mean Absolute Percentage Error (GRG-MAPE). For ANN, the SD is 2.13, the RMSE is 2.11 dB, and the GRG-MAPE is 0.96. Also, for the NFS, the SD is 1.99, the RMSE is 2.06 dB, and the GRG-MAPE is 0.97. It should be noted that the transition zone between the city and the river was characterized by the proposed ANN and NFS in contrast with the classic literature models, which did not demonstrate coherence in the transition zone.

Motivated by applications in wireless communications, in this paper we propose a reconstruction algorithm for sparse signals whose values are taken from a discrete set, using a limited number of noisy observations. Unlike conventional compressed sensing algorithms, the proposed approach incorporates knowledge of the discrete valued nature of the signal in the detection process. This is accomplished through the alternating direction method of the multipliers which is applied as a heuristic to decompose the associated maximum likelihood detection problem in order to find candidate solutions with a low computational complexity order. Numerical results in different scenarios show that the proposed algorithm is capable of achieving very competitive recovery error rates when compared with other existing suboptimal approaches.

Federated Multi-Agent Deep Reinforcement Learning (FMADRL) is gathering keen research interests, as it may offer efficient solutions towards meeting the extreme requirements of future wireless communication networks. By contrast to centralized Deep Reinforcement Learning (DRL) and Multi-Agent DRL (MADRL), F-MADRL enables edge devices to cooperate without sharing their private data, while reducing the delays and signaling costs inherent to centralized approaches. In this article, we explore the new opportunities brought by F-MADRL with a special focus on optimized wireless communications for Internet of Things (IoT) applications. Firstly, we categorize state-of-theart F-MADRL approaches, based on some distinctive features such as aggregation frequency, signaling overhead and privacy level. To better illustrate the behavior and advantages of FMADRL, it is numerically compared to its fully centralized and distributed independent DRL counterparts, through a Sub- 6GHz/mmWave band association optimization problem for IoT wireless communications. Finally, we identify and discuss the open research directions and challenges, in order to spur further interest in this promising area. [ABSTRACT FROM AUTHOR]

Emotion recognition systems are typically trained to classify a given psychophysiological state into emotion categories. Current platforms for emotion ground-truth collection show limitations for real-world scenarios of long-duration content (e.g. $> $ > 10 minutes), namely: 1) Real-time annotation tools are distracting and become exhausting; 2) Perform retrospective annotation of the whole content in bulk (providing highly coarse annotations); or 3) Are used by external experts (depending on the number of annotators and their subjective experience). We explore a novel approach, the EmotiphAI Annotator, that allows undisturbed content visualisation and simplifies the annotation process by using segmentation algorithms that select brief clips for emotional annotation retrospectively. We compare three methods for content segmentation based on physiological data (Electrodermal Activity (EDA), emotion-based), scene (time-based), and random (control) selection. The EmotiphAI Annotator attained a B+ System Usability Scale score and low-average mental workload as per the NASA Task Load Index (40%). The reliability of the self-report was analysed by the inter-rater agreement (STD $< $ < 0.75), coherence across time segmentation methods (STD $< $ < 0.17), comparison against the state-of-the-art ground truth (STD $< $ < 0.7), and correlation to EDA ($> $ > 0.3 to 0.8), where the EDA-based method obtained the overall best performance. [ABSTRACT FROM AUTHOR]