10784 research outputs found
Sort by
Bioaccessibility and Antioxidant Capacity of Grape Seed and Grape Skin Phenolic Compounds After Simulated In Vitro Gastrointestinal Digestion
Get PDFPublisher Copyright: © The Author(s) 2024.Grapes present recognized beneficial effects on human health due to their polyphenolic composition. The grape overproduction together with the wine sales down and the world socioeconomic situation makes the wine grape valorization a promising strategy to give an added-value to this natural product. The objective of the present work was to study the influence of in vitro gastrointestinal digestion on antioxidant capacity and polyphenolic profile of skin and seed extracts of different grape varieties (Tempranillo, Graciano, Maturana tinta and Hondarrabi zuri). After in vitro gastrointestinal digestion, total phenolic content (TPC) of seed polyphenolic extracts decreased significantly for all the varieties. The highest decrease was for Tempranillo going from 108 ± 9 to 50 ± 3 mg / g dry matter (dm). This variety also showed the highest decrease of 90% in antioxidant capacity. However, for all the skin polyphenolic extracts there was an increase in TPC. The highest variation was also for Tempranillo. It varied from 10.1 ± 0.8 to 55.1 ± 0.9 mg / g dm. Among red varieties Tempranillo skin polyphenolic extract showed the lowest undigested anthocyanin content but the highest bioaccessibility index (BI) of 77%. For flavanols, flavonols and procyanidins the seed polyphenolic extracts showed a BI at the intestinal phase between 11% for (+)-epicatechin gallate to 130% procyanidin A2. The results of this study suggest that grape skin extracts and grape seed extracts are a reliable source of bioaccessible antioxidant polyphenols, to be used for the development of antioxidant supplements with specific functionalities depending on the grape variety.Peer reviewe
The Use of Virtual Sensors for Bead Size Measurements in Wire-Arc Directed Energy Deposition
Get PDFPublisher Copyright: © 2024 by the authors.Having garnered significant attention in the scientific community over the past decade, wire-arc directed energy deposition (arc-DED) technology is at the heart of this investigation into additive manufacturing parameters. Singularly focused on Invar as the selected material, the primary objective revolves around devising a virtual sensor for the indirect size measurement of the bead. This innovative methodology involves the seamless integration of internal signals and sensors, enabling the derivation of crucial measurements sans the requirement for direct physical interaction or conventional measurement methodologies. The internal signals recorded, the comprising voltage, the current, the energy from the welding heat source generator, the wire feed speed from the feeding system, the traverse speed from the machine axes, and the temperature from a pyrometer located in the head were all captured through the control of the machine specially dedicated to the arc-DED process during a phase of optimizing and modeling the bead geometry. Finally, a feedforward neural network (FNN), also known as a multi-layer perceptron (MLP), is designed, with the internal signals serving as the input and the height and width of the bead constituting the output. Remarkably cost-effective, this solution circumvents the need for intricate measurements and significantly contributes to the proper layer-by-layer growth process. Furthermore, a neural network model is implemented with a test loss of 0.144 and a test accuracy of 1.0 in order to predict weld bead geometry based on process parameters, thus offering a promising approach for real-time monitoring and defect detection.Peer reviewe
A novel methodology for day-ahead buildings energy demand forecasting to provide flexibility services in energy markets
Get PDFPublisher Copyright: © 2024In future smart grid environment, local energy markets will become a reality to provide flexibility. Consequently, it will be essential not only to implement accurate energy consumption forecasters at the building level to determine which buildings can provide the required flexibility, but also at an aggregated level to anticipate power system boundary conditions. Thus, both forecasters play a key role in supporting the reliable and secure operation of smart grids and developing future demand response strategies. Although there is a piece of literature that addressed energy demand forecasting for day-ahead horizons, proposed algorithms only focused on improving accuracy neglecting energy markets technical boundary conditions. This study presents a novel methodology based on random forest machine learning algorithm to predict day-ahead energy demand at individual buildings with a 15-minute resolution. Furthermore, an analysis has been conducted to assess whether the application of time-series decomposition techniques or shape factors can enhance the accuracy of the proposed methodology. The results indicate that the proposed methodology is effective and accurate, exhibiting a MAPE of 10.77% – 31.52% and an R2 of 0.51–0.70 for individual buildings. These findings demonstrate the potential of the methodology for future energy markets.This research was conducted as part of OMEGA-X project (https://omega-x.eu), which has received funding from the European Union’s Horizon Europe Framework Programme under Grant Agreement No. 101069287. The authors are grateful for the support and contributions from other members of the OMEGA-X project consortium, particularly to the Municipality of Maia and the University of Maia for providing the electrical consumption data of the buildings and meteorological data.Peer reviewe
Probing perfection: The relentless art of meddling for pulmonary airway segmentation from HRCT via a human-AI collaboration based active learning method
Get PDFPublisher Copyright: © 2024 The AuthorsIn the realm of pulmonary tracheal segmentation, the scarcity of annotated data stands as a prevalent pain point in most medical segmentation endeavors. Concurrently, most Deep Learning (DL) methodologies employed in this domain invariably grapple with other dual challenges: the inherent opacity of ‘black box’ models and the ongoing pursuit of performance enhancement. In response to these intertwined challenges, the core concept of our Human-Computer Interaction (HCI) based learning models (RS_UNet, LC_UNet, UUNet and WD_UNet) hinge on the versatile combination of diverse query strategies and an array of deep learning models. We train four HCI models based on the initial training dataset and sequentially repeat the following steps 1–4: (1) Query Strategy: Our proposed HCI models selects those samples which contribute the most additional representative information when labeled in each iteration of the query strategy (showing the names and sequence numbers of the samples to be annotated). Additionally, in this phase, the model selects the unlabeled samples with the greatest predictive disparity by calculating the Wasserstein Distance, Least Confidence, Entropy Sampling, and Random Sampling. (2) Central line correction: The selected samples in previous stage are then used for domain expert correction of the system-generated tracheal central lines in each training round. (3) Update training dataset: When domain experts are involved in each epoch of the DL model's training iterations, they update the training dataset with greater precision after each epoch, thereby enhancing the trustworthiness of the ‘black box’ DL model and improving the performance of models. (4) Model training: Proposed HCI model is trained using the updated training dataset and an enhanced version of existing UNet. Experimental results validate the effectiveness of this Human-Computer Interaction-based approaches, demonstrating that our proposed WD-UNet, LC-UNet, UUNet, RS-UNet achieve comparable or even superior performance than the state-of-the-art DL models, such as WD-UNet with only 15 %–35 % of the training data, leading to substantial reductions (65 %–85 % reduction of annotation effort) in physician annotation time.Peer reviewe
Large Language Model Operations (LLMOps): Definition, Challenges, and Lifecycle Management
No full textPublisher Copyright: © 2024 University of Split, FESB.Numerous studies explore the prospects presented by the recent upsurge of large language models. The usage of LLMs in production environments poses challenges that highlight the limitations of methodologies such as MLOps, and further investigation in this field is required. To this end, a new methodology, coined large language model operations (LLMOps), has arisen to address the particularities of LLMs. This term is so recent that the scientific literature has not yet agreed on a common definition for it, and the use of non-peer reviewed studies becomes a must. In this research, we review the current literature in the field to shed light on the adoption of LLMOps to drive innovation and efficiency in deploying large language models in real-world applications. To this end, three research questions are used to guide the contribution to the scientific literature with a unified definition of LLMOps, the challenges posed by LLMs that require the need for this new methodology, and to outline the key stages of LLMOps and their particularities that must be considered.Peer reviewe
Wireless Power Transfer for Unmanned Underwater Vehicles: Technologies, Challenges and Applications
Get PDFPublisher Copyright: © 2024 by the authors.Unmanned underwater vehicles (UUVs) are key technologies to conduct preventive inspection and maintenance tasks in offshore renewable energy plants. Making such vehicles autonomous would lead to benefits such as improved availability, cost reduction and carbon emission minimization. However, some technological aspects, including the powering of these devices, remain with a long way to go. In this context, underwater wireless power transfer (UWPT) solutions have potential to overcome UUV powering drawbacks. Considering the relevance of this topic for offshore renewable plants, this work aims to provide a comprehensive summary of the state of the art regarding UPWT technologies. A technology intelligence study is conducted by means of a bibliographical survey. Regarding underwater wireless power transfer, the main methods are reviewed, and it is concluded that inductive wireless power transfer (IWPT) technologies have the most potential. These inductive systems are described, and their challenges in underwater environments are presented. A review of the underwater IWPT experiments and applications is conducted, and innovative solutions are listed. Achieving efficient and reliable UWPT technologies is not trivial, but significant progress is identified. Generally, the latest solutions exhibit efficiencies between 88% and 93% in laboratory settings, with power ratings reaching up to 1–3 kW. Based on the assessment, a power transfer within the range of 1 kW appears to be feasible and may be sufficient to operate small UUVs. However, work-class UUVs require at least a tenfold power increase. Thus, although UPWT has advanced significantly, further research is required to industrially establish these technologies.Peer reviewe
Overview of the Methods and Applications of Electrodermal Activity Assessment and Its Relation with Electrotactile Feedback and Potential Use in Automated Calibration
No full textPublisher Copyright: © 2024 IEEE.This paper presents an overview of the Electrodermal Activity (EDA) applications and methods for signal analysis, as it is a tool that has a wide range of potential uses in monitoring of user's wellbeing. Significant correlations have been shown between EDA and stress, both when induced through external stimulation (auditory, visual, pain, etc.), cognitive load or physiological strain (heat, dehydration, intense physical activity). Also, along with the heartrate it is one of the least intrusive parameters to be measured and requires relatively simple signal processing. All of this makes it a highly promising tool in a number of applications. However, naturally occurring variability between subjects and between different measurement sessions of the same subjects require significant effort in the experimental setup, as well as meticulous calibration procedures, in order to obtain reliable results. This largely limits the use of EDA to highly controlled laboratory or clinical conditions, significantly limiting its real-world impact. There is an observed connection between the EDA responses and electrotactile stimulation which can be highly controlled. This may potentially be used in an automated calibration procedure that can be rapidly performed before each monitoring session and would not require any expertise of the end user. Here we introduce this hypothesis to lay the foundation for the future experimental work.Peer reviewe
Toward Fully Automated Inspection of Critical Assets Supported by Autonomous Mobile Robots, Vision Sensors, and Artificial Intelligence
Get PDFPublisher Copyright: © 2024 by the authors.Robotic inspection is advancing in performance capabilities and is now being considered for industrial applications beyond laboratory experiments. As industries increasingly rely on complex machinery, pipelines, and structures, the need for precise and reliable inspection methods becomes paramount to ensure operational integrity and mitigate risks. AI-assisted autonomous mobile robots offer the potential to automate inspection processes, reduce human error, and provide real-time insights into asset conditions. A primary concern is the necessity to validate the performance of these systems under real-world conditions. While laboratory tests and simulations can provide valuable insights, the true efficacy of AI algorithms and robotic platforms can only be determined through rigorous field testing and validation. This paper aligns with this need by evaluating the performance of one-stage models for object detection in tasks that support and enhance the perception capabilities of autonomous mobile robots. The evaluation addresses both the execution of assigned tasks and the robot’s own navigation. Our benchmark of classification models for robotic inspection considers three real-world transportation and logistics use cases, as well as several generations of the well-known YOLO architecture. The performance results from field tests using real robotic devices equipped with such object detection capabilities are promising, and expose the enormous potential and actionability of autonomous robotic systems for fully automated inspection and maintenance in open-world settings.Peer reviewe
Non-imaging Medical Data Synthesis for Trustworthy AI: A Comprehensive Survey
Get PDFPublisher Copyright: © 2024 Copyright held by the owner/author(s).Data quality is a key factor in the development of trustworthy AI in healthcare. A large volume of curated datasets with controlled confounding factors can improve the accuracy, robustness, and privacy of downstream AI algorithms. However, access to high-quality datasets is limited by the technical difficulties of data acquisition, and large-scale sharing of healthcare data is hindered by strict ethical restrictions. Data synthesis algorithms, which generate data with distributions similar to real clinical data, can serve as a potential solution to address the scarcity of good quality data during the development of trustworthy AI. However, state-of-the-art data synthesis algorithms, especially deep learning algorithms, focus more on imaging data while neglecting the synthesis of non-imaging healthcare data, including clinical measurements, medical signals and waveforms, and electronic healthcare records (EHRs). Therefore, in this article, we will review synthesis algorithms, particularly for non-imaging medical data, with the aim of providing trustworthy AI in this domain. This tutorial-style review article will provide comprehensive descriptions of non-imaging medical data synthesis, covering aspects such as algorithms, evaluations, limitations, and future research directions.Peer reviewe
Sustainable Cultural Tourism: Proposal for a Comparative Indicator-Based Framework in European Destinations
Get PDFPublisher Copyright: © 2024 by the authors.Effective decision-making in tourism destinations relies significantly on employing suitable indicators for policy design and impact evaluation. However, the adoption of sustainability-focused indicators remains constrained in the field of cultural tourism. The purpose of this research is to provide decision-makers with an extensive array of criteria and indicators, enabling informed decision-making, policy formulation, and impact assessment tailored to the distinctive attributes encountered in European destinations. Based on the synthesis of existing approaches, and in co-creation with 21 European tourism destinations, an indicator-based framework is proposed, structured around the environmental, economic, social, cultural, resilience, and characterization domains. The results are particularly novel in the resilience and cultural domains, related to the recovery from crisis impacts, but also to the enhancement of digital approaches, as well as the preservation and promotion of cultural heritage towards a more hospitable destination. Moreover, the involvement of stakeholders incorporating real-case scenarios allows this research to bridge the gap between theoretical constructs and practical application. The indicator-based framework resulting from this research will provide stakeholders with assistance in assessing and comparing the impacts of cultural tourism on their destinations and, thence, help them acquire knowledge on cultural resource management, contributing to a more sustainable, responsible, and balanced impact.Peer reviewe