Fine-grained classification of marine animals supports ecology, biodiversity and habitat conservation, and evidence-based policy-making. However, existing methods often overlook contextual interactions from the surrounding environment and insufficiently incorporate the hierarchical structure of marine biological taxonomy. To address these challenges, we propose MATANet (Multi-context Attention and Taxonomy-Aware Network), a novel model designed for fine-grained marine species classification. MATANet mimics expert strategies by using taxonomy and environmental context to interpret ambiguous features of underwater animals. It consists of two key components: a Multi-Context Environmental Attention Module (MCEAM), which learns relationships between regions of interest (ROIs) and their surrounding environments, and a Hierarchical Separation-Induced Learning Module (HSLM), which encodes taxonomic hierarchy into the feature space. MATANet combines instance and environmental features with taxonomic structure to enhance fine-grained classification. Experiments on the FathomNet2025, FAIR1M, and LifeCLEF2015-Fish datasets demonstrate state-of-the-art performance. The source code is available at: https://github.com/dhlee-work/fathomnet-cvpr2025-ssl
This paper examines the work of two English shipbuilders, William and Francis Warden, who, at the behest of the Portuguese Crown, served as master shipwrights at the royal shipyards in Lisbon. Their contributions were pivotal in shaping the character of Portuguese shipbuilding and naval design during the first half of the eighteenth century, decades before similar developments would be seen in Spain, in what was referred to at the time as the ‘English manner’. Moreover, they played a key role in advancing the efforts aimed at the standardisation of a shipbuilding typology across the Atlantic in Brazil. Given that the activity of English shipbuilders abroad was not subject to the strict regulations of the Royal Navy Establishment's, this paper will both seek to define the new model introduced by the Warden and explore the extent to which the Portuguese Navy's pursuit of uniformity and efficiency was genuinely effective.
This study evaluates the applicability and limitations of the International Maritime Organization (IMO)’s maneuvering standards (MSC.137(76)) for large fishing vessels under 100 m in length, which are not currently included in the regulation. Full-scale turning circle, zig-zag, and stopping tests were conducted on three representative vessels—a stern trawler, a purse seiner, and a squid-jigging boat—in accordance with ISO 15016:2015 and ITTC procedures. All the vessels satisfied the IMO criteria for their turning and stopping performance; however, the zig-zag tests revealed distinct differences in directional stability. The stern trawler and purse seiner showed excessive first-overshoot angles, indicating over-reactive yaw responses influenced by the hull form and propulsion–rudder interaction, whereas the squid-jigging boat exhibited very small overshoot angles, reflecting strong yaw damping. These patterns correspond with variations in the block coefficient (C<sub>b</sub>), Froude number (F<sub>n</sub>), and length-to-breadth ratio L<sub>BP</sub>/B. Although all vessels met the IMO stopping requirements, their deceleration behavior differed due to their hull fullness and reverse-thrust efficiency. Overall, the findings clearly demonstrate a mismatch between merchant-vessel-based IMO standards and the maneuvering characteristics of fishing vessels, which require agility and frequent low-speed operations. The results provide a basis for refining maneuvering prediction methods and developing assessment criteria tailored to fishing vessel design and operational profiles.
ObjectiveIn order to improve the power quality of ships under complex working conditions, an energy management strategy based on wavelet decomposition is proposed for fuel cell ships with a hybrid energy storage system (HESS) consisting of a battery and supercapacitor. Methods First, wavelet decomposition and fuzzy logic control are used to distribute the load power of the ship and optimize the charging and discharging process of the battery. Next, grey wolf optimizer （GWO） is used to optimize the parameters of the HESS and ensure that the energy management strategy matches the equipment parameters. Finally, the ship power system model is built using the Matlab/Simulink platform, and the simulation experiment is verified.ResultsThe simulation results show that the proposed energy management strategy can effectively suppress the output power fluctuation of the fuel cell and realize rational power distribution among devices, reducing the optimized bus voltage fluctuation by 23.19%. ConclusionThis study can provide useful references for the optimization design of the power quality of ships.
This study presents an empirical formula for predicting the ultimate compressive strength of curved plates incorporating welding-induced defects, with the objective of enhancing structural design for Ocean Mobility applications. The proposed formula uniquely considers both initial deflection and welding residual stress, two major sources of imperfection. It introduces the plate slenderness ratio (β) and the flank angle (θ, in radians) as internal variables. It enables the prediction of ultimate strength across eight representative scenarios, defined by combinations of welding direction, loading condition, initial deflection level, and residual stress distribution. The results indicate that welding residual stress can reduce the ultimate strength by up to 10 %, and the proposed formula demonstrates high accuracy with an average deviation within 0.1 % from FEM results. This research improves existing design equations by systematically incorporating the effects of welding defects, and the proposed formula may serve as a reliable tool for accurate ultimate strength assessment in the structural design of welded curved plates.
Alexandra Malyugina, Guoxi Huang, Eduardo Ruiz
et al.
Underwater videos often suffer from degraded quality due to light absorption, scattering, and various noise sources. Among these, marine snow, which is suspended organic particles appearing as bright spots or noise, significantly impacts machine vision tasks, particularly those involving feature matching. Existing methods for removing marine snow are ineffective due to the lack of paired training data. To address this challenge, this paper proposes a novel enhancement framework that introduces a new approach for generating paired datasets from raw underwater videos. The resulting dataset consists of paired images of generated snowy and snow, free underwater videos, enabling supervised training for video enhancement. We describe the dataset creation process, highlight its key characteristics, and demonstrate its effectiveness in enhancing underwater image restoration in the absence of ground truth.
Marine ecosystems face increasing pressure due to climate change, driving the need for scalable, AI-powered monitoring solutions to inform effective conservation and restoration efforts. This paper examines the rapid emergence of underwater AI as a major research frontier and analyzes the factors that have transformed marine perception from a niche application into a catalyst for AI innovation. We identify three convergent drivers: i) environmental necessity for ecosystem-scale monitoring, ii) democratization of underwater datasets through citizen science platforms, and iii) researcher migration from saturated terrestrial computer vision domains. Our analysis reveals how unique underwater challenges - turbidity, cryptic species detection, expert annotation bottlenecks, and cross-ecosystem generalization - are driving fundamental advances in weakly supervised learning, open-set recognition, and robust perception under degraded conditions. We survey emerging trends in datasets, scene understanding and 3D reconstruction, highlighting the paradigm shift from passive observation toward AI-driven, targeted intervention capabilities. The paper demonstrates how underwater constraints are pushing the boundaries of foundation models, self-supervised learning, and perception, with methodological innovations that extend far beyond marine applications to benefit general computer vision, robotics, and environmental monitoring.
This paper deals with the key operational problems of wind turbosets, especially offshore, where vibrations are generated by rotor blades, as a consequence of erosive wear or icing. The primary causes of the imbalance of wind turbine rotors have been characterised, the observable symptoms of which include various forms of vibrations, transmitted from the turbine wheel to the bearing nodes of the power train components. Their identification was the result of an active diagnostic experiment, which actually entered the aerodynamic-mass imbalance of a turbine rotor into a wind power train, built as a small scale model. The recording of the observed monitoring parameters (vibration, aerodynamic, mechanical and electrical) made it possible to determine a set of symptoms (syndrome) of the deteriorated (entered) dynamic state of the entire wind turboset. This provides the basis for positive verification of the assumed concept and methodology of diagnostic testing, the constructed laboratory station and the measuring equipment used. For this reason, testing continued, taking into account the known and recognisable faults that most often occur during the operation of offshore wind turbosets. Transferring the results of this type of model research to full-size, real objects makes it possible to detect secondary (fatigue) damage to the elements transmitting torque from the wind turbine rotor to the generator early, especially the thrust bearings or gear wheel teeth.
Agustín Romero-Vargas, Luis Alberto Fdez-Güelfo, Ana Blandino
et al.
This study focuses on mitigating the socio-economic and environmental damage of the invasive macroalga <i>Rugulopteryx okamurae</i> and counteracting the pollution from petroleum-based plastics by using the alga as a feedstock for polyhydroxybutyrate (PHB) production. The enzymatic hydrolysis of <i>R. okamurae,</i> non-pretreated and hydrothermally acid-pretreated (0.2 N HCl, 15 min), was carried out, reaching reducing sugar (RS) concentrations of 10.7 g/L and 21.7 g/L, respectively. The hydrolysates obtained were used as a culture medium for PHB production with <i>Cupriavidus necator,</i> a Gram-negative soil bacterium, without supplementation with any external carbon and nitrogen sources. The highest yield (0.774 g PHB/g RS) and biopolymer accumulation percentage (89.8% cell dry weight, CDW) were achieved with hydrolysates from pretreated macroalga, reaching values comparable to the highest reported in the literature. Hence, it can be concluded that hydrolysates obtained from algal biomass hydrothermally pretreated with acid have a concentration of sugars and a C/N ratio that favour PHB production.
In the presence of dynamic uncertainties, external time-varying disturbances, and limited inputs to the multi-point mooring system (MPMS) of a floating offshore platform (FOP), this paper proposes a robust adaptive dynamic surface (RADS) control method incorporating a disturbance observer. A disturbance observer is designed to estimate the unknown time-varying disturbance and apply feedforward compensation to the control variable. Simultaneously, the adaptive law of the σ-corrected leakage term is employed to estimate the bound of the disturbance observation error, thereby enhancing positioning accuracy. An auxiliary dynamic system (ADS) is then introduced to address input constraints, while the differential explosion problem associated with the traditional inversion method is resolved through the integration of the dynamic surface control (DSC) algorithm. The Lyapunov function is utilized to demonstrate that the controller ensures the consistent ultimate boundedness of all signals within the closed-loop system. Finally, a simulation experiment was conducted based on the eight-point mooring platform of the “Kantan3”, and the positioning accuracy reached 3%, which is higher than the specification requirements of the classification society. The results indicate that the designed controller achieves higher positioning accuracy and improved anti-interference performance and has been put into practical application on “Kantan3”.
Underwater optical sensor networks offer significant advantages over traditional acoustic sensor networks, such as high bandwidth and low latency, making them a research hotspot to realize high-speed and real-time transmission of underwater data in sea-air cross-domain communication. However, traditional routing protocols for underwater sensor networks often encounter problems such as routing holes and redundant data transmission due to the transmission obstruction and directional alignment of underwater optical nodes. To address these issues, a fuzzy logic-based opportunistic routing protocol for underwater optical sensor networks was proposed. Firstly, a fuzzy logic-based algorithm was proposed to evaluate the underwater optical channel links. Combined with the modeling of an underwater complex communication environment, the multi-factor fusion evaluation of routing metrics was designed. Additionally, a dynamic mechanism was designed to set the node data forwarding probability and data packet retention time based on real-time link evaluation data. Finally, a probabilistic redundant suppression forwarding algorithm was proposed. Simulation results demonstrate that the proposed routing protocol effectively improves transmission efficiency and reduces end-to-end latency in typical marine communication environments, exhibiting good network dynamic adaptability.
This paper presents an experimental and numerical analysis of the response of a scaled double-bottom structure with high and penetrated girders and floors impacted vertically by a rock-shaped indenter. The specimen, scaled from the bottom structure of the power-battery cabin of a new energy ship, is struck by a spherical indenter. The special double-bottom structure is designed to protect the power batteries and to facilitate heat dissipation. The experimental overall impact response, vibration acceleration, and stress of the inner bottom plate are measured in order to evaluate the impact environment in the target cabin. The investigation provides valuable information to evaluate the safety of power-battery cabins in a ship grounding scenario. The experimental results show good agreement with the finite element analyses using the explicit LS-DYNA software. The numerical analysis outlines the influence of the structural openings on the impact response and also the effect of battery mass and striking velocity on the impact environment in the target cabin.
With the increasing popularity of recommendation systems (RecSys), the demand for compute resources in datacenters has surged. However, the model-wise resource allocation employed in current RecSys model serving architectures falls short in effectively utilizing resources, leading to sub-optimal total cost of ownership. We propose ElasticRec, a model serving architecture for RecSys providing resource elasticity and high memory efficiency. ElasticRec is based on a microservice-based software architecture for fine-grained resource allocation, tailored to the heterogeneous resource demands of RecSys. Additionally, ElasticRec achieves high memory efficiency via our utility-based resource allocation. Overall, ElasticRec achieves an average 3.3x reduction in memory allocation size and 8.1x increase in memory utility, resulting in an average 1.6x reduction in deployment cost compared to state-of-the-art RecSys inference serving system.
The Pearl River Delta (PRD, China) has undergone complex geological development within a multi-island faulted basin, shaped by the interplay of regional tectonic movements, Quaternary sea-level fluctuations, and fluvial-marine interactions. Despite a great number of studies on the Holocene sedimentary sequences and spatial differences of lithofacies and environments, scant attention has been paid to the overarching human influence on deltaic evolution and coastline modifications since the Neolithic epoch. To further elucidate the spatial variation in Holocene sedimentation and its underlying basement topography shaped during the Last Glacial Maximum (LGM), we compiled a comprehensive dataset incorporating borehole data from over 2800 cores (the maximum depth can reach 92.5 m) within the PRD. Subsequently, high-resolution isobath maps of Quaternary deltaic deposits were generated, offering unprecedented insights into sediment distribution. This dataset facilitated a nuanced reconstruction of pre-Holocene topography, revealing a zone characterized by elongated, deep-incised valleys governed by NW-SE fault orientations. Further, we delineated coastline shifts since the period of maximum Holocene transgression (~7000 years BP), contributing to an enhanced understanding of the formation and evolutionary patterns of the delta and river network oscillations. Our findings illuminate an increasing anthropogenic impact on the rate of fluvial sedimentation and land growth, particularly accentuated over the last two millennia, favoring deltaic accretion.
The prevailing offshore field development solutions, i.e., dry tree and wet tree systems, are confronted with serious technical and economic challenges in deep and ultra-deep waters resulting from the large depth of water, far offshore distance, and harsh ocean environmental conditions, as well as high cost. In response to these challenges, an innovative Deepwater Artificial Seabed (DAS) production system is proposed in this article. The DAS production system concentrates on well access and riser design, which enables shallow-water-rated subsea production systems to develop Deepwater (DW) and Ultra-Deepwater (UDW) fields. First, DW & UDW field development drivers are discussed and presented. This is followed by a detailed discussion of the merits and demerits of the prevailing dry tree and wet tree field development solutions. On this basis, the design philosophy and main characteristics of the DAS production system are presented and discussed in detail. Dynamic survival analysis for the fully coupled Floating Production Storage and Offloading (FPSO)-DAS production system is carried out. The artificial seabed stability is systematically investigated for both intact and damaged conditions. The global analysis results indicate that the DAS production system as developed experiences quasi-static responses even under extreme storm conditions, due to the location of the artificial seabed and the decoupling effects of the flexible jumpers. The new DAS production system is considered to be a competitive and cost-effective field development solution in depths of up to 3000 m.
Fatemeh Parikhani, Ehsan Atazadeh, Jafar Razeghi
et al.
This work is the first in a series, and its purpose is the comprehensive assessment of the ecological state of the Aras River using biological indicators of water quality by diatoms based on species’ ecological preferences, pollution indices, statistics, and ecological mapping. Samples of diatoms and soft algae and measurements of water quality were analyzed at sixteen sampling sites (between 2020 and 2022) along the Aras River. The impact of anthropological activity on the river was monitored concerning water quality, river health, and ecosystem function. The physical and chemical characteristics of the water were measured. The biological properties of the algal periphyton communities, including species composition, were also measured. Based on the studies conducted in this research, 280 species were identified. The most prosperous species were <i>Diatoma vulgaris</i>, <i>Amphora ovalis</i>, <i>Cocconeis placentula</i>, <i>Rhoicosphenia abbre-viatae</i>, <i>Cymbella helvetica</i>, <i>Brevisira arentii</i>, <i>Navicula tripunctata</i>, <i>Nitzschia linearis</i>, <i>Microcystis botrys</i>, <i>Microcystis aeruginosa</i>, <i>Pseudanabaena limnetica</i>, <i>Scenedesmus obliquus</i>, and <i>Pleurosira laevis</i> (a pollution-resistant and salinity-resistant species first found in aquatic habitats in the Aras River). As a result, the empirical data and algal indices showed the river’s lower reaches to be in poor condition. Exploration of the algal assemblage and water chemistry data using computationally unconstrained ordination techniques such as principal component analysis (PCA) and canonical correspondence analysis (CCA) indicated two strong gradients in the data sets. The results support that water body classification is a function of water chemistry and biological and hydrological characteristics, as it is necessary to include pollutant effects on biota since the nature of the receiving waters influences the river’s water quality.
Osip Kokin, Irina Usyagina, Nikita Meshcheriakov
et al.
Ice scours are formed when the keels of floating icebergs or sea ice hummocks penetrate unlithified seabed sediments. Until now, ice scours have been divided into “relict” and “modern” according to the water depth that corresponds with the possible maximum vertical dimensions of the keels of modern floating icebergs. However, this approach does not consider climatic changes at the present sea level, which affect the maximum depth of ice keels. We present an application of <sup>210</sup>Pb dating of the largest ice scour in the Baydaratskaya Bay area (Kara Sea), located at depths of about 28–32 m. Two sediment cores were studied; these were taken on 2 November 2021 from the R/V <i>Akademik Nikolay Strakhov</i> directly in the ice scour and on the “background” seabed surface, not processed via ice scouring. According to the results of <sup>210</sup>Pb dating, the studied ice scour was formed no later than the end of the Little Ice Age. Based on the extrapolation of possible sedimentation rates prior to 1917 (0.22–0.38 cm/year), the age of the ice scour is estimated to be 1810 ± 30 AD. The mean rate of ice scour filling with 70 cm thick sediments from the moment of its formation is around 0.33 cm/year.
In Time-Triggered (TT) or time-sensitive networks, the transmission of a TT frame is required to be scheduled at a precise time instant for industrial distributed real-time control systems. Other (or {\em best-effort} (BE)) frames are forwarded in a BE manner. Under this scheduling strategy, the transmission of a TT frame must wait until its scheduled instant even if it could have been transmitted sooner. On the other hand, BE frames are transmitted whenever possible but may miss deadlines or may even be dropped due to congestion. As a result, TT transmission and BE delivery are incompatible with each other. To remedy this incompatibility, we propose a synergistic switch architecture (SWA) for TT transmission with BE delivery to dynamically improve the end-to-end (e2e) latency of TT frames by opportunistically exploiting BE delivery. Given a TT frame, the SWA generates and transmits a cloned copy with BE delivery. The first frame arriving at the receiver device is delivered with a configured jitter and the other copy ignored. So, the SWA achieves shorter latency and controllable jitter, the best of both worlds. We have implemented SWA using FPGAs in an industry-strength TT switches and used four test scenarios to demonstrate SWA's improvements of e2e latency and controllable jitter over the state-of-the-art TT transmission scheme.
Noise: an enemy to be dealt with and a major factor limiting communication system performance. However, what if there is gold in that garbage? In conventional engineering, our focus is primarily on eliminating, suppressing, combating, or even ignoring noise and its detrimental impacts. Conversely, could we exploit it similarly to biology, which utilizes noise-alike carrier signals to convey information? In this context, the utilization of noise, or noise-alike signals in general, has been put forward as a means to realize unconditionally secure communication systems in the future. In this tutorial article, we begin by tracing the origins of thermal noise-based communication and highlighting one of its significant applications for ensuring unconditionally secure networks: the Kirchhoff-law-Johnson-noise (KLJN) secure key exchange scheme. We then delve into the inherent challenges tied to secure communication and discuss the imperative need for physics-based key distribution schemes in pursuit of unconditional security. Concurrently, we provide a concise overview of quantum key distribution (QKD) schemes and draw comparisons with their KLJN-based counterparts. Finally, extending beyond wired communication loops, we explore the transmission of noise signals over-the-air and evaluate their potential for stealth and secure wireless communication systems.
Knowledge engineering is the process of creating and maintaining knowledge-producing systems. Throughout the history of computer science and AI, knowledge engineering workflows have been widely used because high-quality knowledge is assumed to be crucial for reliable intelligent agents. However, the landscape of knowledge engineering has changed, presenting four challenges: unaddressed stakeholder requirements, mismatched technologies, adoption barriers for new organizations, and misalignment with software engineering practices. In this paper, we propose to address these challenges by developing a reference architecture using a mainstream software methodology. By studying the requirements of different stakeholders and eras, we identify 23 essential quality attributes for evaluating reference architectures. We assess three candidate architectures from recent literature based on these attributes. Finally, we discuss the next steps towards a comprehensive reference architecture, including prioritizing quality attributes, integrating components with complementary strengths, and supporting missing socio-technical requirements. As this endeavor requires a collaborative effort, we invite all knowledge engineering researchers and practitioners to join us.