Abstract
Water bodies are complex and interconnected systems that play a crucial role in both our environment and our economy. Studying these water bodies is therefore essential but collecting and analyzing water samples can be challenging, particularly when dealing with large volumes of water. This article presents a system capable of autonomously filtering large volumes of water through standard marine biology filters and preserving them to later be analyzed. The system's preliminary results are presented in this paper.

Abstract
In recent years, Artificial Intelligence (AI) provided essential tools to enhance the productivity of activities related to civil engineering, particularly in design, construction, and maintenance. In this framework, the present work proposes a novel AI computer vision methodology for automatically identifying the corrosion phenomenon on roofing systems of large-scale industrial buildings. The proposed method can be incorporated into computational packages for easier integration by the industry to enhance the inspection activities' performance. For this purpose, a dedicated image database with more than 8k high-resolution aerial images was developed for supervised training. An Unmanned Aerial Vehicle (UAV) was used to acquire remote georeferenced images safely and efficiently. The corrosion anomalies were manually annotated using a segmentation strategy summing up 18,381 instances. These anomalies were identified through instance segmentation using the Mask based Region-Convolution Neural Network (Mask R-CNN) framework adjusted to the created dataset. Some adjustments were performed to enhance the performance of the classification model, particularly defining an adequate input image size, data augmentation strategy, Intersection over a Union (IoU) threshold during training, and type of backbone network. The inferences show promising results, with correct detections even under complex backgrounds, poor illumination conditions, and instances of significantly reduced dimensions. Furthermore, in scenarios without a roofing system, the model proved reliable, not producing any false positive occurrences. The best model achieved metrics' values equal to 65.1% for the bounding box detection Average Precision (AP) and 59.2% for the mask AP, considering an IoU of 50%. Regarding classification metrics, the precision and recall were equal to 85.8% and 84.0%, respectively. The developed methodology proved to be extremely valuable for guiding infrastructure managers in taking physically informed decisions based on the real assets condition.

Abstract
Greenland ice core records display abrupt transitions, designated as Dansgaard-Oeschger (DO) events, characterised by episodes of rapid warming (typically decades) followed by a slower cooling. The identification of abrupt transitions is hindered by the typical low resolution and small size of paleoclimate records, and their significant temporal variability. Furthermore, the amplitude and duration of the DO events varies substantially along the last glacial period, which further hinders the objective identification of abrupt transitions from ice core records Automatic, purely data-driven methods, have the potential to foster the identification of abrupt transitions in palaeoclimate time series in an objective way, complementing the traditional identification of transitions by visual inspection of the time series.In this study we apply an algorithmic time series method, the Matrix Profile approach, to the analysis of the NGRIP Greenland ice core record, focusing on:- the ability of the method to retrieve in an automatic way abrupt transitions, by comparing the anomalies identified by the matrix profile method with the expert-based identification of DO events;- the characterisation of DO events, by classifying DO events in terms of shape and identifying events with similar warming/cooling temporal patternThe results for the NGRIP time series show that the matrix profile approach struggles to retrieve all the abrupt transitions that are identified by experts as DO events, the main limitation arising from the diversity in length of DO events and the method’s dependence on fixed-size sub-sequences within the time series. However, the matrix profile method is able to characterise the similarity of shape patterns between DO events in an objective and consistent way.

Abstract
<p>Gamma radiation measured over the ocean is mainly due to airborne radionuclides, as gamma emission by radon degassing from the ocean is negligible. Airborne gamma-emitting elements include radon progeny (Pb-2114, Bi-214, Pb-210) and cosmogenic radionuclides such as Be-7. Radon progeny attaches readily to aerosols, thus the fate of gamma-emitting radon progeny, after its formation by radioactive decay from radon, is expected to be closely linked to that of aerosols.</p> <p>Gamma radiation measurements over the Atlantic Ocean were made on board the ship-rigged sailing ship NRP Sagres in the framework of project SAIL (Space-Atmosphere-Ocean Interactions in the marine boundary Layer). The measurements were performed continuously with a NaI(Tl) scintillator counting all gamma rays from 475 keV to 3 MeV.  </p> <p>The counts from the sensor were recorded every 1 second into a computer system which had his time reference corrected by a GNSS pulse per second (PPS) signal. The GNSS was also used to precisely position the ship. The measurements were performed over the Atlantic ocean from January to May 2020, along the ship’s round trip from Lisboa - Cape Verde – Rio de Janeiro – Buenos Aires – Cape Town – Cape Verde - Lisboa.</p> <p>The results show that the gamma radiation time series displays considerable higher counts and larger variability in January compared to the remaining period. Reanalysis data also indicate higher aerosol concentration. This work investigates in detail the association between the temporal evolution of the gamma radiation measurements obtained from the SAIL campaign over the Atlantic Ocean and co-located total aerosol concentration at 550 nm obtained every 3 hours from EAC4(ECMWF Atmospheric Composition Reanalysis 4) data.</p>

Abstract
We present MONET, a new multimodal dataset captured using a thermal camera mounted on a drone that flew over rural areas, and recorded human and vehicle activities. We captured MONET to study the problem of object localisation and behaviour understanding of targets undergoing large-scale variations and being recorded from different and moving viewpoints. Target activities occur in two different land sites, each with unique scene structures and cluttered backgrounds. MONET consists of approximately 53K images featuring 162K manually annotated bounding boxes. Each image is timestamp-aligned with drone metadata that includes information about attitudes, speed, altitude, and GPS coordinates. MONET is different from previous thermal drone datasets because it features multimodal data, including rural scenes captured with thermal cameras containing both person and vehicle targets, along with trajectory information and metadata. We assessed the difficulty of the dataset in terms of transfer learning between the two sites and evaluated nine object detection algorithms to identify the open challenges associated with this type of data. Project page: https://github.com/fabiopoiesi/monet-dataset.

Abstract
Wiki articles are created and maintained by a crowd of editors, producing a continuous stream of reviews. Reviews can take the form of additions, reverts, or both. This crowdsourcing model is exposed to manipulation since neither reviews nor editors are automatically screened and purged. To protect articles against vandalism or damage, the stream of reviews can be mined to classify reviews and profile editors in real-time. The goal of this work is to anticipate and explain which reviews to revert. This way, editors are informed why their edits will be reverted. The proposed method employs stream-based processing, updating the profiling and classification models on each incoming event. The profiling uses side and content-based features employing Natural Language Processing, and editor profiles are incrementally updated based on their reviews. Since the proposed method relies on self-explainable classification algorithms, it is possible to understand why a review has been classified as a revert or a non-revert. In addition, this work contributes an algorithm for generating synthetic data for class balancing, making the final classification fairer. The proposed online method was tested with a real data set from Wikivoyage, which was balanced through the aforementioned synthetic data generation. The results attained near-90% values for all evaluation metrics (accuracy, precision, recall, and F-measure).