Earthquakes are sudden releases of energy in the Earth's crust that generate seismic waves recorded as time‐series signals called seismograms. Earthquake analysis depends on interpreting the temporal structure of seismic phases, yet many machine-learning approaches require large manually labeled datasets that are difficult to obtain at scale. To address this, we propose a self-supervised framework that learns waveform continuity and ordering cues without human annotations. Our approach consists of two steps: preprocessing and model training. In preprocessing, each one-dimensional waveform is converted into a standardized waveform image and partitioned into a fixed number of consecutive segments with small gaps to simulate missing observations and create non-contiguous inputs. In model training, these segments are randomly shuffled and used to train a Siamese convolutional neural network to predict the shuffled order through a permutation-classification objective, where pseudo-labels are generated automatically from the applied shuffle. We evaluate the framework across multiple segment settings and summarize its performance trends as task difficulty increases. Overall, this study demonstrates that permutation-based self-supervision can enable learning useful waveform structure from unlabeled data, and it suggests a practical direction for developing phase-aware representations that can support future seismic analysis when labeled data are limited.

To address the nonlinear and non-stationary characteristics of vibration signals during rolling bearing operation and the issue of insufficient degradation information representation, this paper designs a prediction framework that combines Adaptive Variational Mode Decomposition (AVMD) with a SETCN-BiGRU multi-head temporal attention mechanism for Remaining Useful Life (RUL) prediction. First, AVMD is used to decompose the bearing horizontal vibration signal into five Intrinsic Mode Functions (IMFs). Time-domain and frequency-domain statistics are extracted from each IMF and concatenated into a 115-dimensional degradation feature sequence. Subsequently, the model processes in parallel: a TCN-SENet branch extracts local temporal features and adaptively adjusts channel weights, while a BiGRU with multi-head temporal attention sub-network captures global bidirectional dependencies and critical degradation periods within the degradation sequence. Finally, the two types of features are fused, and the RUL prediction result is output. Experimental results demonstrate that the proposed model achieves an RMSE, MAE, and R² of 0.0582, 0.0477, and 0.9483 respectively on the IEEE PHM 2012 dataset, and average values of 0.0780, 0.0559, and 0.9133 on a self-built laboratory bearing dataset, indicating good prediction accuracy, robustness, and generalization ability.

This study investigates how accelerated sea-level rise interacts with tidal dynamics along the East Coast of the United States, which are regions influenced by the Gulf Stream. Using 2004–2019 satellite altimetry, NOAA tide-gauge data, and the Atlantic Meridional Overturning Circulation (AMOC) index, the analysis reveals that both sea-surface height (SSH) and tidal range (TR) exhibit consistent increases. Although their year-to-year variations are weakly linked, SSH and TR share a strong long-term slope change. We speculate that increases in SSH may deepen the water and are associated with increased TR, although the underlying mechanisms remain uncertain. Integrating AMOC reanalysis data indicates that the weakening of its circulation likely drives the SSH rise and tidal changes. A potential regime transition around 2020 introduces new uncertainty, suggesting the AMOC system may be entering a different dynamical state. These findings highlight the need for improved multi-decadal modeling to support coastal resilience and future ocean renewable energy station planning.

This study proposes a multi-attribute, multi-task Transformer model framework to address structural health monitoring (SHM) and the optimization of sustainable bio-asphalt formulations. The φ-NeXt dataset, comprising 37,000 multi-label structural damage images, was constructed to enable simultaneous multi-task detection for visual SHM. The model was then transferred to the bio-asphalt domain through prompt fine-tuning, thereby mitigating the challenge of data scarcity. Furthermore, by integrating Gaussian processes grounded in physical mechanisms, Bayesian optimization, and reinforcement learning, I established a multi-objective formulation optimization workflow and obtained Pareto-optimal formulations that meet engineering requirements. Experimental results demonstrate that the proposed method performs abouAEI-26-205t 95% accuracy in multi-task detection and long-term performance prediction, providing theoretical support for the application of eco-asphalt materials in road engineering.

Fully solid-state batteries, with their inherent high safety and high energy density, are emerging as a core area of development for next-generation electrochemical energy storage technologies. However, cathode materials remain a critical bottleneck in determining their overall performance. Iron-based cathode materials offer advantages such as abundant resources, low cost, environmental friendliness and flexible lithium storage mechanisms, and have seen a series of breakthroughs in the field of all-solid-state batteries in recent years. This review systematically examines the electrochemical reaction mechanisms, performance characteristics and modification strategies of lithium iron phosphate-based insertion materials, halide-based mixed-conducting materials, and sulfide and oxide-based transition materials. It points out that poor solid–solid interface compatibility, slow reaction kinetics and significant volume effects during charging and discharging are common core challenges across these material types, with interface engineering, integrated electrode design and self-healing mechanisms representing key paths to overcoming these bottlenecks. Lastly, looking to the future, the industrialisation of iron-based cathode materials requires a focus on multi-scale co-design, as well as breakthroughs in low-cost, large-scale fabrication processes and full-cell integration technologies, thereby providing the core foundation for the commercial application of high-safety, low-cost all-solid-state batteries.

At present, artificial intelligence technology can achieve simple communication with users and provide certain answers to the questions raised by users. With the vigorous development of artificial intelligence technology, obtaining emotional companionship through artificial intelligence technology has become a trend and trend. This study focuses on the application of multimodal artificial intelligence technology in emotional companionship, especially in providing personalized emotional companionship. It summarizes relevant literature and research results in recent years, and summarizes the current situation and problems of artificial intelligence technology in achieving personalized emotional companionship under multimodal conditions. It provides reference for solving problems such as difficulty in analyzing and extracting information from historical information. Research has found that at present, artificial intelligence technology can better understand user expressions and provide positive feedback in personalized emotional companionship, but there are still drawbacks such as "forgetting memories", "awkward conversations", and insufficient interaction effects.

This study investigates the persistence of distinct national tactical identities across Europe's top five football leagues (Premier League, La Liga, Serie A, Bundesliga, and Ligue 1) amidst the backdrop of increasing globalization. Utilizing a longitudinal dataset of 588 team-season observations derived from FIFA attributes (2010–2016), the analysis employs cluster analysis, principal component analysis, and classification modeling to examine stylistic differences and temporal evolution. The findings confirm that while tactical diffusion is evident, systematic structural differences persist: the Bundesliga and Serie A are characterized by high tempo and physical aggression, La Liga maintains a slower, possession-oriented approach, and Ligue 1 exhibits distinct preferences for structured buildup play. The classification model successfully predicts league affiliation with a macro-AUC of 0.84, demonstrating that tactical attributes contain sufficient information to distinguish leagues above random chance. Although temporal analysis indicates a slight trend toward convergence, the results suggest that complete homogenization has not occurred, as long-standing competitive environments and cultural traditions continue to sustain recognizable league-level stylistic identities.

In modern communications, optical fiber communication has become an indispensable means of information transmission. With the rapid development of cloud computing, the Internet, short-video platforms, artificial intelligence, and related technologies, the demand for higher transmission rates and broader bandwidth in optical fiber communication systems has continued to grow. As fiber transmission rates increase, the constraints imposed by nonlinear effects on the development of optical fiber communications have become increasingly significant. One typical nonlinear effect is four-wave mixing (FWM), which has a substantial impact on high-power and high-speed optical communication systems. This paper introduces three types of nonlinear effect noise, namely Rayleigh scattering, stimulated Brillouin scattering, and four-wave mixing. Particular emphasis is placed on the four-wave mixing effect. To suppress FWM, a method combining thin-film filters with three effective channel allocation schemes is proposed, including non-uniform channel spacing, enlarged channel intervals, and increased guard bandwidth. The proposed approach contributes to the suppression of four-wave mixing noise in high-speed optical communication systems and significantly improves system reliability and stability, especially in transmission lines operating under heavy traffic loads, such as dense wavelength division multiplexing (DWDM) systems.

Against the background of China's dual-carbon goals and refined energy management on the building side, university dormitory buildings have become typical scenarios for the coordinated application of rooftop photovoltaics and electrochemical energy storage. This paper takes Dormitory Building No. 9 at the University of Science and Technology Beijing, constructing a grid-connected microgrid with rooftop PV, lithium iron phosphate storage, loads, and the grid. Using Beijing's time-of-use industrial/commercial tariff and typical summer/winter days, a 24-hour linear optimal scheduling model minimizes daily operating cost (power purchase cost plus PV curtailment penalty) under constraints like power balance, SOC, charge/discharge limits, and grid interaction. The results show that, under the boundary condition of no power export to the grid, the energy storage system reconstructs the grid purchase curve through "off-peak charging, high-price discharging, and midday absorption of surplus PV power." The daily operating costs in the typical summer and winter scenarios decrease by 16.41% and 18.59%, respectively; PV curtailment is reduced to zero in both scenarios; and the PV self-consumption rate increases to 100%. Sensitivity analysis indicates that increasing the storage capacity can further reduce operating cost, but the marginal benefit declines gradually, providing a quantitative reference for sizing energy storage in dormitory buildings.