Search Results (79)

Accurate photovoltaic (PV) power forecasting is critical for the stable operation of power systems. Existing methods rely solely on historical data, which significantly decline in forecasting accuracy at 3–4 h ahead. To address this problem, a novel ultra-short-term PV power forecasting method based on temporal attention-variable parallel fusion encoder network is proposed to enhance the stability of forecasting results by incorporating Numerical Weather Prediction data to correct temporal predictions. Specifically, independent encoding modules are constructed for both historical power sequences and future NWP sequences, enabling deep feature extraction of their respective temporal characteristics. During the decoding phase, a two-stage coupled decoding strategy is employed: for 1–8 steps predictions, the model relies solely on temporal features, while for 9–16 steps horizons, it dynamically fuses encoded information from historical power data and future NWP inputs. This approach allows for accurate characterization of future trend dynamics. Experimental results demonstrate that, compared with conventional methods, the proposed model reduces the average normalized root mean square error (NRMSE) at 4th ultra-short-term forecasting by 0.50–5.20%, while it improves the $R^2$ by 0.047–0.362, validating the effectiveness of the proposed approach. Full article

The phenylpropanoid compounds are crucial secondary metabolites for blueberry plants. Low temperatures induce the expression of phenylpropanoid biosynthesis genes and regulate the accumulation of phenylpropanoid metabolites. However, the molecular mechanisms of blueberry leaves in response to low-temperature stress are unknown. To explore the molecular mechanisms of phenylpropanoid biosynthesis under low-temperature stress, the 6-month-old blueberry plants were cultured at 10 °C for 0, 6, 12, 24, and 48 h. The total of 16,388 differentially expressed genes (DEGs) and 303 differentially accumulated metabolites (DAMs) were identified by transcriptome deep sequencing (RNA-seq) and ultra-high performance liquid mass spectrometry, respectively. The most enriched low-temperature-responsive genes are mainly involved in the phenylpropanoid biosynthesis pathway and the main low-temperature-responsive metabolites come from the phenylpropanoid superclass based on transcriptome and metabolome data, respectively. CBF2 plays essential roles in the ICE-CBF-COR regulatory pathway, and transcription factors (TFs) ERF109, MYB14, WRKY40, HSP30, MPSR1, ZHD4, MADS3, and MADS27 might be responsible for blueberry leaf low-temperature tolerance. The MYB TFs from group 5, group 6, and group AtMYB5 may regulate the accumulation of phenylpropanoid metabolites by regulating expression of phenylpropanoid biosynthesis genes. These findings uncover possible molecular mechanisms of phenylpropanoid biosynthesis during low-temperature stress and provide a basis for future studies and crop improvement. Full article

The Extra-Long Weir Construction method for deep foundation pit construction is crucial for urban underground development. However, as excavation projects become deeper and more complex, construction safety risks increase markedly. Existing monitoring technologies and numerical simulation models face persistent challenges: high uncertainty in risk occurrence, complex environmental interactions, and difficulties in extracting effective warning signals from multi-source data. To address these challenges, this study establishes a systematic risk evaluation framework comprising 6 primary and 29 secondary indicators through Fault Tree Analysis and develops a novel DL-MSD (Deep Learning and Multi-Source Data Prediction) model integrating CNN, ResUnit, and LSTM networks for spatiotemporal sequence analysis and multi-source data fusion. Validated using 6524 samples from the Jinji Lake Tunnel project, the model employs single-factor prediction for hazard source tracing and multi-factor fusion for comprehensive risk assessment. Results demonstrate exceptional performance: 90.2% average accuracy for single-factor warnings and 77.1% for multi-factor fusion, with, critically, all severe warnings (Level I risks) identified with zero omissions. Comparative analysis with T-S fuzzy neural networks, EWT-NARX, and Random Forest confirmed superior accuracy and computational efficiency. An integrated platform incorporating BIM and IoT technologies enables automated monitoring, intelligent prediction, and adaptive control. This study establishes a data-driven intelligent early warning framework that significantly improves prediction accuracy, timeliness, and reliability in deep foundation pit construction, marking a paradigm shift from reactive response to proactive prevention. The findings provide theoretical and methodological support for safety management in ultra-deep excavation projects, offering reliable decision-making evidence for enhancing construction safety and risk management. Full article

The transition to 5G networks brings unprecedented speed, ultra-low latency, and massive connectivity. Nevertheless, it introduces complex traffic patterns and broader attack surfaces that render traditional intrusion detection systems (IDSs) ineffective. Existing rule-based methods and classical machine learning approaches struggle to capture the temporal and dynamic characteristics of 5G traffic, while many deep learning models lack interpretability, making them unsuitable for high-stakes security environments. To address these challenges, we propose Bidirectional Temporal Anomaly Detector (BiTAD), a deep temporal learning architecture for anomaly detection in 5G networks. BiTAD leverages dual-direction temporal sequence modelling with attention to encode both past and future dependencies while focusing on critical segments within network sequences. Like many deep models, BiTAD’s faces interpretability challenges. To resolve its “black-box” nature, a dual-perspective explainability module, coined TwinLens, is proposed. This module integrates SHAP and TimeSHAP to provide global feature attribution and temporal relevance, delivering dual-perspective interpretability. Evaluated on the public 5G-NIDD dataset, BiTAD demonstrates superior detection performance compared to existing models. TwinLens enables transparent insights by identifying which features and when they were most influential to anomaly predictions. By jointly addressing the limitations in temporal modelling and interpretability, our work contributes a practical IDS framework tailored to the demands of next-generation mobile networks. Full article

Deep-well radar telemetry over ultra-long cables suffers from strong frequency-selective attenuation and impedance drift under high temperature and pressure. We have proposed a channel-adaptive “communication + acquisition” architecture for a 7500 m borehole radar system. The scheme integrates spread-spectrum time domain reflectometry (SSTDR; m-sequence with BPSK) to monitor the cable in situ, identify termination/cable impedance, and adaptively match the load, thereby reducing reflection-induced loss. On the receiving side, we combine time domain adaptive equalization—implemented as an LMS-driven FIR filter—with frequency domain OFDM equalization based on least-squares (LS) channel estimation, enabling constellation recovery and robust demodulation over the distorted channel. The full processing chain is realized in real time on a Xilinx Artix-7 (XC7A100T) FPGA with module-level reuse and pre-stored training sequences for efficient hardware scheduling. In a field deployment in the Shunbei area at 7500 m depth, radar results show high agreement with third-party geological logs: the GR-curve correlation reaches 0.92, the casing reflector at ~7250 m is clearly reproduced, and the key bottom depth error is 0.013%. These results verify that the proposed system maintains stable communication and accurate imaging in harsh deep-well environments while remaining compact and implementable on cost-effective hardware. Full article

With the rise of 6G, the exponential growth of data traffic, the proliferation of emerging applications, and the ubiquity of smart devices, the demand for spectral resources is unprecedented. Terahertz communication (100 GHz–3 THz) plays a key role in alleviating spectrum scarcity through ultra-broadband transmission. In this study, terahertz optical carrier-based systems are employed, where fiber-optic components are used to generate the optical signals, and the signal is transmitted via direct detection in the receiver side, without relying on fiber-optic transmission. In these systems, deep learning-based equalization effectively compensates for nonlinear distortions, while probability shaping (PS) enhances system capacity under modulation constraints. However, the probability distribution of signals processed by PS varies with amplitude, making it challenging to extract useful information from the minority class, which in turn limits the effectiveness of nonlinear equalization. Furthermore, in IM-DD systems, optical multipath interference (MPI) noise introduces signal-dependent amplitude jitter after direct detection, degrading system performance. To address these challenges, we propose a lightweight neural network equalizer assisted by the Synthetic Minority Oversampling Technique (SMOTE) and a clustering method. Applying SMOTE prior to the equalizer mitigates training difficulties arising from class imbalance, while the low-complexity clustering algorithm after the equalizer identifies edge jitter levels for decision-making. This joint approach compensates for both nonlinear distortion and jitter-related decision errors. Based on this algorithm, we conducted a 3.75 Gbaud W-band PAM4 wireless transmission experiment over 300 m at Fudan University’s Handan campus, achieving a bit error rate of 1.32 × 10⁻³, which corresponds to a 70.7% improvement over conventional schemes. Compared to traditional equalizers, the proposed new equalizer reduces algorithm complexity by 70.6% and training sequence length by 33%, while achieving the same performance. These advantages highlight its significant potential for future optical carrier-based wireless communication systems. Full article

The advancement of high-throughput sequencing has revolutionised genomic research by generating large amounts of data. However, Whole-Genome Sequencing is associated with a statistical challenge known as the p >> n problem. We classified 1825 individuals into five breeds based on 11,915,233 SNPs. First, three feature selection algorithms were applied: SNP-tagging and two approaches based on supervised rank aggregation, followed by either one-dimensional (1D-SRA) or multidimensional (MD-SRA) feature clustering. Individuals were then classified into breeds using a deep learning classifier composed of Convolutional Neural Networks. SNPs selected by SNP-tagging yielded the least satisfactory F1-score (86.87%); however, this approach offered rapid computing time. The 1D-SRA was less suitable for ultra-high-dimensional data due to computational, memory, and storage limitations. However, the SNP set selected by this algorithm provided the best classification quality (96.81%). MD-SRA provided a good balance between classification quality (95.12%) and computational efficiency (17x lower analysis time, 14x lower data storage). Unlike SNP-tagging, SRA-based approaches are universal and are not limited to genomic data. This study addressed the demand for efficient computational and statistical tools for feature selection in high-dimensional genomic data. The results demonstrate that the proposed MD-SRA is suitable for the classification of high-dimensional data. Full article

This study proposes a hybrid approach to analyze and forecast non-stationary financial time series by combining statistical models with deep neural networks. A model is introduced that integrates three key components: the Generalized Autoregressive Score (GAS) model, which captures volatility dynamics; an attention mechanism (ATT), which identifies the most relevant features within the sequence; and a Long Short-Term Memory (LSTM) neural network, which receives the outputs of the previous modules to generate price forecasts. This architecture is referred to as GAS-ATT-LSTM. Both unidirectional and bidirectional variants were evaluated using real financial data from the Nasdaq Composite Index, Invesco QQQ Trust, ProShares UltraPro QQQ, Bitcoin, and gold and silver futures. The proposed model’s performance was compared against five benchmark architectures: LSTM Bidirectional, GARCH-LSTM Bidirectional, ATT-LSTM, GAS-LSTM, and GAS-LSTM Bidirectional, under sliding windows of 3, 5, and 7 days. The results show that GAS-ATT-LSTM, particularly in its bidirectional form, consistently outperforms the benchmark models across most assets and forecasting horizons. It stands out for its adaptability to varying volatility levels and temporal structures, achieving significant improvements in both accuracy and stability. These findings confirm the effectiveness of the proposed hybrid model as a robust tool for forecasting complex financial time series. Full article

Underwater sonar target recognition is crucial in fields such as national defense, navigation, and environmental monitoring. However, it faces issues such as the complex characteristics of ship-radiated noise, imbalanced data distribution, non-stationarity, and bottlenecks of existing technologies. This paper proposes the MultiFuseNet-AID network, aiming to address these challenges. The network includes the TriFusion block module, the novel lightweight attention residual network (NLARN), the long- and short-term attention (LSTA) module, and the Mamba module. Through the TriFusion block module, the original, differential, and cumulative signals are processed in parallel, and features such as MFCC, CQT, and Fbank are fused to achieve deep multi-domain feature fusion, thereby enhancing the signal representation ability. The NLARN was optimized based on the ResNet architecture, with the SE attention mechanism embedded. Combined with the long- and short-term attention (LSTA) and the Mamba module, it could capture long-sequence dependencies with an O(N) complexity, completing the optimization of lightweight long sequence modeling. At the same time, with the help of feature fusion, and layer normalization and residual connections of the Mamba module, the adaptability of the model in complex scenarios with imbalanced data and strong noise was enhanced. On the DeepShip and ShipsEar datasets, the recognition rates of this model reached 98.39% and 99.77%, respectively. The number of parameters and the number of floating point operations were significantly lower than those of classical models, and it showed good stability and generalization ability under different sample label ratios. The research shows that the MultiFuseNet-AID network effectively broke through the bottlenecks of existing technologies. However, there is still room for improvement in terms of adaptability to extreme underwater environments, training efficiency, and adaptability to ultra-small devices. It provides a new direction for the development of underwater sonar target recognition technology. Full article

Over the past decade, microbial natural products research has witnessed a transformative “deep-mining era” driven by key technological advances such as high-throughput sequencing (e.g., PacBio HiFi), ultra-sensitive HRMS (resolution ≥ 100,000), and multi-omics synergy. These innovations have shifted discovery from serendipitous isolation to data-driven, targeted mining. These innovations have transitioned discovery from serendipitous isolation to data-driven targeted mining. Genome mining pipelines (e.g., antiSMASH 7.0 and DeepBGC) can now systematically discover hidden biosynthetic gene clusters (BGCs), especially in under-explored taxa. Metabolomics has achieved unprecedented accuracy, enabling researchers to target novel compounds in complex extracts. Integrated strategies—combining genomic prediction, metabolomics analysis, and experimental validation—constitute new paradigms of current “deep mining”. This review provides a systematic overview of 185 novel microbial natural products discovered between 2018 and 2024, and dissects how these technological leaps have reshaped the discovery paradigm from traditional isolation to data-driven mining. Full article

This study systematically analyzes reservoir formation mechanisms under deep burial conditions, integrating macroscopic observations from representative ultra-deep clastic reservoirs in four major sedimentary basins in central and western China. Developing effective clastic reservoirs in ultra-deep strata (6000–8000 m) remains a critical yet debated topic in petroleum geology. Recent advances in exploration techniques and geological understanding have challenged conventional views, confirming the presence of viable clastic reservoirs at such depths. Findings reveal that reservoir quality in ultra-deep strata is preserved and enhanced through the interplay of sedimentary, diagenetic, and tectonic processes. Key controlling factors include (1) high-energy depositional environments promoting primary porosity development, (2) proximity to hydrocarbon source rocks enabling multi-phase hydrocarbon charging, (3) overpressure and low geothermal gradients reducing cementation and compaction, and (4) late-stage tectonic fracturing that significantly improves permeability. Additionally, dissolution porosity and fracture networks formed during diagenetic and tectonic evolution collectively enhance reservoir potential. The identification of favorable reservoir zones under the sedimentation–diagenesis-tectonics model provides critical insights for future hydrocarbon exploration in ultra-deep clastic sequences. Full article

Deep neural video compression codecs have shown great promise in recent years. However, there are still considerable challenges for ultra-low bitrate video coding. Inspired by recent diffusion models for image and video compression attempts, we attempt to leverage diffusion models for ultra-low bitrate portrait video compression. In this paper, we propose a predictive portrait video compression method that leverages the temporal prediction capabilities of diffusion models. Specifically, we develop a temporal diffusion predictor based on a conditional latent diffusion model, with the predicted results serving as decoded frames. We symmetrically integrate a temporal diffusion predictor at the encoding and decoding side, respectively. When the perceptual quality of the predicted results in encoding end falls below a predefined threshold, a new frame sequence is employed for prediction. While the predictor at the decoding side directly generates predicted frames as reconstruction based on the evaluation results. This symmetry ensures that the prediction frames generated at the decoding end are consistent with those at the encoding end. We also design an adaptive coding strategy that incorporates frame quality assessment and adaptive keyframe control. To ensure consistent quality of subsequent predicted frames and achieve high perceptual reconstruction, this strategy dynamically evaluates the visual quality of the predicted results during encoding, retains the predicted frames that meet the quality threshold, and adaptively adjusts the length of the keyframe sequence based on motion complexity. The experimental results demonstrate that, compared with the traditional video codecs and other popular methods, the proposed scheme provides superior compression performance at ultra-low bitrates while maintaining competitiveness in visual effects, achieving more than 24% bitrate savings compared with VVC in terms of perceptual distortion. Full article

Effective analysis of central bank communications is critical for anticipating monetary policy changes and guiding market expectations. However, traditional natural language processing models face significant challenges in processing lengthy and nuanced policy documents, which often exceed tens of thousands of tokens. This study addresses these challenges by proposing a novel integrated deep learning framework based on Hyena Hierarchy architectures, which utilize sub-quadratic convolution mechanisms to efficiently process ultra-long sequences. The framework employs Delta-LoRA (low-rank adaptation) for parameter-efficient fine-tuning, updating less than 1% of the total parameters without additional inference overhead. To ensure robust performance across institutions and policy cycles, domain-adversarial neural networks are incorporated to learn domain-invariant representations, and a multi-task learning approach integrates auxiliary hawkish/dovish sentiment signals. Evaluations conducted on a comprehensive dataset comprising Federal Open Market Committee statements and European Central Bank speeches from 1977 to 2024 demonstrate state-of-the-art performance, achieving over 6% improvement in macro-F1 score compared to baseline models while significantly reducing inference latency by 65%. This work offers a powerful and efficient new paradigm for handling ultra-long financial policy texts and demonstrates the effectiveness of integrating advanced sequence modeling, efficient fine-tuning, and domain adaptation techniques for extracting timely economic signals, with the aim to open new avenues for quantitative policy analysis and financial market forecasting. Full article

Soil porosity, as an essential indicator for assessing soil quality, plays a key role in guiding agricultural production, so it is beneficial to detect soil porosity. However, the currently available methods do not apply to high-precision and rapid detection of soil with a black-box nature in the field, so this paper proposes a soil porosity detection method based on ultrasound and multi-scale CNN-LSTM. Firstly, a series of ring cutter soil samples with different porosities were prepared manually to simulate soil collected in the field using a ring cutter, followed by ultrasonic signal acquisition of the soil samples. The acquired signals were subjected to three kinds of data augmentation processes to enrich the dataset: adding Gaussian white noise, time shift transformation, and random perturbation. Since the collected ultrasonic signals belong to long-time series data and there are different frequency and sequence features, this study constructs a multi-scale CNN-LSTM deep neural network model using large convolution kernels based on the idea of multi-scale feature extraction, which uses multiple large convolution kernels of different sizes to downsize the collected ultra-long time series data and extract local features in the sequences, and combining the ability of LSTM to capture global and long-term dependent features enhances the feature expression ability of the model. The multi-head self-attention mechanism is added at the end of the model to infer the before-and-after relationship of the sequence data to improve the degradation of the model performance caused by waveform distortion. Finally, the model was trained, validated, and tested using ultrasonic signal data collected from soil samples to demonstrate the accuracy of the detection method. The model has a coefficient of determination of 0.9990 for detecting soil porosity, with a percentage root mean square error of only 0.66%. It outperforms other advanced comparative models, making it very promising for application. Full article

Accurate Total Organic Carbon (TOC) prediction in the deeply buried Lower Cambrian Qiongzhusi Formation shale is constrained by extreme heterogeneity (TOC variability: 0.5–12 wt.%, mineral composition Coefficient of Variation > 40%) and ambiguous geophysical responses. This study introduces three key innovations to address these challenges: (1) A Dynamic Weighting–Calibrated Random Forest Regression (DW-RFR) model integrating high-resolution Gamma-Ray-guided dynamic time warping (±0.06 m depth alignment precision derived from 237 core-log calibration points using cross-validation), Principal Component Analysis-Deyang–Anyue Rift Trough Shapley Additive Explanations (PCA-SHAP) hybrid feature engineering (89.3% cumulative variance, VIF < 4), and Bayesian-optimized ensemble learning; (2) systematic benchmarking against conventional ΔlogR (R² = 0.700, RMSE = 0.264) and multi-attribute joint inversion (R² = 0.734, RMSE = 0.213) methods, demonstrating superior accuracy (R² = 0.917, RMSE = 0.171); (3) identification of Gamma Ray (r = 0.82) and bulk density (r = −0.76) as principal TOC predictors, contrasted with resistivity’s thermal maturity-dependent signal attenuation (r = 0.32 at Ro > 3.0%). The methodology establishes a transferable framework for organic-rich shale evaluation, directly applicable to the Longmaxi Formation and global Precambrian–Cambrian transition sequences. Future directions emphasize real-time drilling data integration and quantum computing-enhanced modeling for ultra-deep shale systems, advancing predictive capabilities in tectonically complex basins. Full article