Search Results (9)

Recent advancements in space exploration have significantly increased the volume of astronomical data, heightening the demand for efficient analytical methods. Concurrently, the considerable energy consumption of machine learning (ML) has fostered the emergence of Green AI, emphasizing sustainable, energy-efficient computational practices. We introduce the first large-scale Green AI benchmark for galaxy morphology classification, evaluating over 30 machine learning architectures (classical, ensemble, deep, and hybrid) on CPU and GPU platforms using a balanced subset of the Galaxy Zoo dataset. Beyond traditional metrics (precision, recall, and F1-score), we quantify inference latency, energy consumption, and carbon-equivalent emissions to derive an integrated EcoScore that captures the trade-off between predictive performance and environmental impact. Our results reveal that a GPU-optimized multilayer perceptron achieves state-of-the-art accuracy of 98% while emitting 20× less CO₂ than ensemble forests, which—despite comparable accuracy—incur substantially higher energy costs. We demonstrate that hardware–algorithm co-design, model sparsification, and careful hyperparameter tuning can reduce carbon footprints by over 90% with negligible loss in classification quality. These findings provide actionable guidelines for deploying energy-efficient, high-fidelity models in both ground-based data centers and onboard space observatories, paving the way for truly sustainable, large-scale astronomical data analysis. Full article

As human and fully automatic live captioning methods coexist and compete against one another, quality analyses and certification become essential. A case in point is LiRICS, the Live Respeaking International Certification Standard created by the Galician Observatory for Media Accessibility (GALMA) to help maintain high international standards in the live captioning profession. Until now, this certification had only been used to assess human captioners. In this paper, it is applied for the first time to automatic captioning (more specifically to Lexi, the automatic software used by the leading captioning company AI-Media) in order to ascertain whether automatic captions have reached an accuracy level that can match that of human captions. After presenting the materials and the methods (NER model), the paper reports on the results of the analysis of Lexi’s English and Spanish automatic captions. With average accuracy rates of 98.56% in English and 98.26% in Spanish, these captions often manage to reach human levels of quality, except when applied to colloquial content featuring several speakers. A final discussion is devoted to a reflection on how automatic and human live captions can coexist as long as the different purposes they serve are considered, namely the access in bulk provided by automatic captions and the curated access offered by human captions. Full article

Visitor attraction dynamics lead tourism industry paths. A complex artificial neural network model was built to predict the incoming tourism flow in the Apulia region of Southern Italy as a function of the heterogeneity of the tourism supply available in this area. Open data from the Regional Tourism Observatory were targeted. Information on the distribution of facilities and activities that attract regional tourist flows was collected and grouped by municipality. An artificial neural network model was built with total tourist attendance as the dependent variable and tourist attractions as regressors. The Root Mean Square Error (RMSE) was used to select the optimal model using the lowest value. The final model was run with a hidden layer consisting of three neurons and a decay value of 0.01. A Multi-Objective Counterfactual model (MOC) was then constructed using a randomly selected row of normalized data frame to validate a useful tool in increasing total tourist attendance by 20% over that of the randomly selected municipality. A Garson’s variables importance plot indicated natural landscapes such as beaches, sea caves, and natural parks have a primary role expressed in terms of variable importance in the AI algorithm when used as an innovative methodology for evaluating tourism flows in the Apulia region. A further MOC model built using a randomly selected row of normalized data frame showed convents, libraries, historical buildings, public gardens, and museums as the top five features most modified to improve total attendance in a randomly selected municipality. Use of AI modeling revealed that the implementation of nature-based solutions may speed up the flow of tourism in the Apulia region while also promoting sustainable social development. Full article

With increasing air and sea temperatures, the thermodynamic environments over the oceans are becoming more favourable for the development of intense tropical cyclones (TCs) with rapid intensification (RI). The South China coastal region consists of highly densely populated cities, especially over the Pearl River Delta (PRD) region. Intense TCs maintaining their strength or the RI of TCs close to the coastal region can present substantial forecasting challenges and have significant potential impacts on the coastal population. This study investigates the effect of sea-surface and sub-surface temperatures and salinity on the intensification of five TCs, namely Super Typhoon Hato in 2017, Super Typhoon Mangkhut in 2018, and Typhoon Talim, Super Typhoon Saola, and Severe Typhoon Koinu in 2023, which have significantly affected the South China coastal region and triggered high TC warning signals in Hong Kong in the past few years. This analysis utilised the Hong Kong Observatory’s TC best-track and intensity data, along with sea temperature and salinity profiles generated using the China Ocean ReAnalysis version 2 (CORA2) product from the National Marine Data and Information Service of China. It was found that high sea-surface temperatures (SST) of 30 °C or above for a depth of about 20 m, low sea-surface salinity (SSS) levels of 33.8 psu or below for a depth of at least 20 m, and strong salinity stratification of at least 0.6 psu per 100 m depth might offer useful hints for predicting the RI of TCs over the western North Pacific and the South China Sea (SCS) in operational forecasting, while noting other contributing environmental factors and synoptic flow patterns conducive to RI. This study represents the first documentation of sub-surface salinity’s impact on some intense TCs traversing the SCS during 2017–2023 based on an observational study. Our aim is to supplement operational techniques for forecasting RI with some quantitative guidance based on upper-level ocean observations of temperatures and salinity, on top of well-known but more rapidly changing dynamical factors like low-level convergence, weak vertical wind shear, and upper-level divergent outflow, as forecasted with numerical weather prediction models. This study will also encourage further research to refine the analysis of quantitative contributions from different RI factors and the identification of essential features for developing AI models as one way to improve the forecasting of TC RI before the TC makes landfall near the PRD, with due consideration given to the effect of freshwater river discharge from the Pearl River. Full article

Through the use of Underwater Smart Sensor Networks (USSNs), Marine Observatories (MOs) provide continuous ocean monitoring. Deployed sensors may not perform as intended due to the heterogeneity of USSN devices’ hardware and software when combined with the Internet. Hence, USSNs are regarded as complex distributed systems. As such, USSN designers will encounter challenges throughout the design phase related to time, complexity, sharing diverse domain experiences (viewpoints), and ensuring optimal performance for the deployed USSNs. Accordingly, during the USSN development and deployment phases, a few Underwater Environmental Constraints (UECs) should be taken into account. These constraints may include the salinity level and the operational depth of every physical component (sensor, server, etc.) that will be utilized throughout the duration of the USSN information systems’ development and implementation. To this end, in this article we present how we integrated an Artificial Intelligence (AI) Database, an extended ArchiMO meta-model, and a design tool into our previously proposed Enterprise Architecture Framework. This addition proposes adding new Underwater Environmental Constraints (UECs) to the AI Database, which is accessed by USSN designers when they define models, with the goal of simplifying the USSN design activity. This serves as the basis for generating a new version of our ArchiMO design tool that includes the UECs. To illustrate our proposal, we use the newly generated ArchiMO to create a model in the MO domain. Furthermore, we use our self-developed domain-specific model compiler to produce the relevant simulation code. Throughout the design phase, our approach contributes to the handling and controling of the uncertainties and variances of the provided quality of service that may occur during the performance of the USSNs, as well as reducing the design activity’s complexity and time. It provides a way to share the different viewpoints of the designers in the domain of USSNs. Full article

Drought has significant impacts on both society and the environment, but it is a gradual and comprehensive process that affects a region over time. Therefore, non-structural measures are necessary to prepare and respond to the damage caused by drought in a flexible manner according to the stage of drought. In this study, an AI-based water demand prediction model was developed using deep neural network (DNN) and long short-term memory (LSTM) models. The model was trained from 2004 to 2015 and verified from 2016 to 2021. Model accuracy was evaluated using data, with the LSTM model achieving a correlation coefficient (CC) of 0.95 and normalized root mean square error (NRMSE) of 8.38, indicating excellent performance. The probability of the random variable X falling within the interval [a,b], as described by the probability density function f(x), was calculated using the water demand data. The cumulative distribution function was used to calculate the probability of the random variable being less than or equal to a specific value. These calculations were used to establish the criteria for each stage of the crisis alert system. Decision tree (DT) and random forest (RF) models, based on AI-based classification, were used to predict water demand at the Gurye intake station. The models took into account the impact of water demand from the previous day, as well as the effects of rainfall, maximum temperature, and average temperature. Daily water demand data from the Gurye intake station and the previous day’s rainfall, maximum temperature, and average temperature data from a nearby observatory were collected from 2004 to 2021. The models were trained on data from 2004 to 2015 and validated on data from 2016 to 2021. Model accuracy was evaluated using the F1-score, with the random forest model achieving a score of 0.88, indicating excellent performance. Full article

The analysis of the 2002 Prestige tanker accident showed how public misinformation can worsen marine pollution incidents, rendering their management suboptimal as these evolve, thus raising the issue of appropriately informing and educating coastal and island populations who are at risk. Two decades later, developments in electronic platforms, including Geographic Information Systems (GIS), the Automatic Identification System (AIS) for ship signal transmission, and social media, provide a set of means for public monitoring of such incidents, creating the possibility to antagonise effectively erroneous or malevolent information, which can hinder efficient actions for containing marine pollution risks even without active training of the populations concerned. The authors, in the framework of the development of the Marine Coastal Observatory and Risk Management project “AEGIS+”, have developed E-S.A.V.E., an online innovative platform that (a) meets the needs of different users as revealed by a survey run across groups of them, (b) uses a suitable Geographic Information System (GIS) environment, (c) cooperates with public authorities, for the reliable update of automated systems, and (d) employs an artificial intelligence (AI)-supported tool for social media monitoring; the platform also provides educational resources and information on national and international resources on marine environmental protection and sustainable maritime logistics. Full article

The irruption of artificial intelligence (AI) and automated technology has substantially changed the journalistic profession, transforming the way of capturing, processing, generating, and distributing information; empowering the work of journalists by modifying the routines and knowledge required by information professionals. This study, which conceptualizes the “exo journalism” on the basis of the impact of AI on the journalism industry, is part of a research project of the Observatory for Information Innovation in the Digital Society (OI2). The results, derived from documentary research supported by case studies and in-depth interviews, propose that AI is a source of innovation and personalization of journalistic content and that it can contribute to the improvement of professional practice, allowing the emergence of a kind of "exo journalist", a conceptual proposal that connects the possibilities of AI with the needs of journalism’s own productive routines. The end result is the enhancement of the journalist’s skills and the improvement of the news product. The research focuses on conceptualizing a kind of support and complement for journalists in the performance of their tasks based on the possibilities of AI in the automatic generation of content and data verification. Full article

In the last years, artificial intelligence (AI) safety gained international recognition in the light of heterogeneous safety-critical and ethical issues that risk overshadowing the broad beneficial impacts of AI. In this context, the implementation of AI observatory endeavors represents one key research direction. This paper motivates the need for an inherently transdisciplinary AI observatory approach integrating diverse retrospective and counterfactual views. We delineate aims and limitations while providing hands-on-advice utilizing concrete practical examples. Distinguishing between unintentionally and intentionally triggered AI risks with diverse socio-psycho-technological impacts, we exemplify a retrospective descriptive analysis followed by a retrospective counterfactual risk analysis. Building on these AI observatory tools, we present near-term transdisciplinary guidelines for AI safety. As further contribution, we discuss differentiated and tailored long-term directions through the lens of two disparate modern AI safety paradigms. For simplicity, we refer to these two different paradigms with the terms artificial stupidity (AS) and eternal creativity (EC) respectively. While both AS and EC acknowledge the need for a hybrid cognitive-affective approach to AI safety and overlap with regard to many short-term considerations, they differ fundamentally in the nature of multiple envisaged long-term solution patterns. By compiling relevant underlying contradistinctions, we aim to provide future-oriented incentives for constructive dialectics in practical and theoretical AI safety research. Full article