Search Results (3,174)

Time-domain signal models have been widely applied to single-channel music source separation tasks due to their ability to overcome the limitations of fixed spectral representations and phase information loss. However, the high acoustic similarity and synchronous temporal evolution between vocals and accompaniment make accurate separation challenging for existing time-domain models. These challenges are mainly reflected in two aspects: (1) the lack of a dynamic mechanism to evaluate the contribution of each source during feature fusion, and (2) difficulty in capturing fine-grained temporal details, often resulting in local artifacts in the output. To address these issues, we propose an attention-driven time-domain convolutional network for vocal and accompaniment source separation. Specifically, we design an embedding attention module to perform adaptive source weighting, enabling the network to emphasize components more relevant to the target mask during training. In addition, an efficient convolutional block attention module is developed to enhance local feature extraction. This module integrates an efficient channel attention mechanism based on one-dimensional convolution while preserving spatial attention, thereby improving the ability to learn discriminative features from the target audio. Comprehensive evaluations on public music datasets demonstrate the effectiveness of the proposed model and its significant improvements over existing approaches. Full article

Chitin, one of the most abundant natural polysaccharides, has gained increasing attention for its structural diversity and potential in biomedicine, agriculture, food packaging, and advanced materials. Conventional chitin production from crustacean shell waste faces limitations, including seasonal availability, allergenic protein contamination, heavy metal residues, and environmentally harmful demineralization processes. Chitin from fungi and microalgae provides a sustainable and chemically versatile alternative. Fungal chitin, generally present in the α-polymorph, is embedded in a chitin–glucan–protein matrix that ensures high crystallinity, mechanical stability, and compatibility for biomedical applications. Microalgal β-chitin, particularly from diatoms, is secreted as high-aspect-ratio microrods and nanofibrils with parallel chain packing, providing enhanced reactivity and structural integrity that are highly attractive for functional materials. Recent progress in green extraction technologies, including enzymatic treatments, ionic liquids, and deep eutectic solvents, enables the recovery of chitin with reduced environmental burden while preserving its native morphology. By integrating sustainable sources with environmentally friendly processing methods, fungal and microalgal chitin offer unique structural polymorphs and tunable properties, positioning them as a promising alternative to crustacean-derived chitin. Full article

User preference mining uses rating data, item content or comments to learn additional knowledge to support the prediction task. For the use of rating data, the usual approach is to take rating matrix as data source, and collaborative filtering as the algorithm to predict user preferences. Item content and comments are usually used in sentiment analysis or as auxiliary information for other algorithms. However, factors such as data sparsity, category diversity, and numerical processing requirements for aspect sentiment analysis affect model performance. This paper proposes a hybrid method, which uses the deep neural network as the basic structure, considers the complementarity of text and numeric data, and integrates the numeric and text embedding into the model. In the construction of text-based embedding, extracts the text summary of each text-based review, and uses the Doc2vec to convert the text summary into multi-dimensional vector. Experiments on two Amazon product datasets show that the proposed model consistently outperforms other baseline models, achieving an average reduction of 15.72% in RMSE, 24.13% in MAE, and 28.91% in MSE. These results confirm the effectiveness of our proposed method for learning user preferences. Full article

In recent years, advances in medicine have been evident thanks to technological growth and interdisciplinary research, which has allowed the integration of knowledge, for example, of engineering into medical fields. This integration has generated developments and new methods that can be applied in alternative situations, highlighting, for example, aspects related to post-stroke therapies, Multiple Sclerosis (MS), or Spinal Cord Injury (SCI) treatments. One of the methods that has stood out and is gaining more acceptance every day is Brain–Computer Interfaces (BCIs), through the acquisition and processing of brain electrical activity, researchers, doctors, and scientists manage to transform this activity into control signals. In turn, there are several methods for operating a BCI, this work will focus on motor imagery (MI)-based BCI and three types of acquisition paradigms (traditional arrow, picture, and video), seeking to improve the accuracy in the classification of motor imagination tasks for naive subjects, which correspond to a MI task for both the left and the right hand. A pipeline and methodology were implemented using the CAR+CSP algorithm to extract the features and simple standard and widely used models such as LDA and SVM for classification. The methodology was tested with post-stroke (PS) subject data with BCI experience, obtaining 96.25% accuracy for the best performance, and with the novel paradigm proposed for the naive subjects, 97.5% was obtained. Several statistical tests were carried out in order to find differences between paradigms within the collected data. In conclusion, it was found that the classification accuracy could be improved by using different strategies in the acquisition stage. Full article

Model-based approaches provide increasingly advanced opportunities for optimizing and accelerating bioprocess development. However, to accurately capture the complexity of biotechnological processes, continuous refinement of suitable models remains essential. A crucial gap in this field has been the lack of suitable model for describing Escherichia coli growth in cultivation media containing yeast extract, while accounting for key bioprocess parameters such as biomass, substrate, acetate, and oxygen. To address this, a published mechanistic macro-kinetic model for E. coli was extended with a set of mathematical equations that describe key aspects of the uptake of yeast extract. The underlying macro-kinetic approach is based on the utilization of amino acids in E. coli, where growth is primarily influenced by two distinct classes of amino acids. Using fed-batch cultivation data from an E. coli K-12 strain supplemented with yeast extract, it was demonstrated that the proposed model extensions were essential for accurately representing the bioprocess. This approach was further validated through fitting the model on cultivation data from five different yeast extracts sourced from various manufacturers. Additionally, the model enabled reliable predictions of growth dynamics across a range of yeast extract concentrations up to 20 g L⁻¹. Further differentiation of the data into batch and fed-batch revealed that for less complex datasets, such as those obtained from a batch phase, a simplified model can be sufficient. Due to its modular structure, the developed model provides the necessary flexibility to serve as a tool for the development, optimization, and control of E. coli cultivations with and without yeast extract. Full article

Background: The aim of the present study is to compare the potential of a high-fat diet, a high-carbohydrate diet, and a high-fat, high-carbohydrate diet to induce liver injury and visceral obesity within a period of five weeks, identify the pattern and degree of hepatic changes at the tissue level, identify the earliest metabolic markers of specific liver changes induced by each type of diet, and to test the possible beneficial effects of steviol glycosides in a rat experimental model. Methods: Wistar rats (n = 56) were divided into seven groups as follows: group BD (before diet), group SD (standard diet), group HFD (high-fat diet), group HCHD (high-carbohydrate diet), group HFHCHD (high-fat high-carbohydrate diet), group SDS (standard diet supplemented with Stevia extract), and group HFDS (high-fat diet supplemented with Stevia extract). Results: Total cholesterol concentrations (2.02 ± 0.22 mmol/L) increased in the HFD group (2.56 ± 0.82 mmol/L) and in the HFDS group (2.89 ± 0.48 mmol/L). The VLDL values before diets were 0.27 ± 0.11 mmol/L and increased most significantly in the HFHCHD group—1.14 ± 0.62 mmol/L. The baseline ALT values (88.4 ± 10.6 U/L) increased in the HFD group (128.13 ± 19.5 U/L) and the HFDS group (127.00 ± 17.74 U/L). Similar increases were registered in the AST/ALT ratio and ALP. Total bilirubin (7.10 ± 1.39 μmol/L) increased in HFD group (27.86 ± 17.01 μmol/L). Serum NO had the lowest values in groups fed diets supplemented with steviol glycosides. All high-calorie diets induced hepatocellular injury. The mass of the perirenal fat depot and cross-sectional area of adipocytes were highest in HFD, HFHCHD, and HFDS groups. Conclusion: High-calorie diets have the potential to induce visceral obesity and hepatocellular injury within a very short period of time, which produces characteristic histological changes and specific biochemical profile. Steviol glycosides may alleviate some aspects of the inflammatory response, but findings about lipid profile parameters and liver enzymes are controversial. Full article

Autonomous commercial vehicles often mount multiple LiDARs to enlarge their field of view, but conventional calibration is labor-intensive and prone to drift during long-term operation. We present an online self-calibration method that combines a ground plane motion constraint with a virtual RGB–D projection, mapping 3D point clouds to 2D feature/depth images to reduce feature extraction cost while preserving 3D structure. Motion consistency across consecutive frames enables a reduced-dimension hand–eye formulation. Within this formulation, the estimation integrates geometric constraints on $SE\left(3\right)$ using Lagrange multiplier aggregation and quasi-Newton refinement. This approach highlights key aspects of identifiability, conditioning, and convergence. An online monitor evaluates plane alignment and LiDAR–INS odometry consistency to detect degradation and trigger recalibration. Tests on a commercial vehicle with six LiDARs and on nuScenes demonstrate accuracy comparable to offline, target-based methods while supporting practical online use. On the vehicle, maximum errors are 6.058 cm (translation) and 4.768° (rotation); on nuScenes, 2.916 cm and 5.386°. The approach streamlines calibration, enables online monitoring, and remains robust in real-world settings. Full article

In recent years, chrono-urbanism has ushered in the x-minute city concept. Effectively combined with the life unit concept, it introduced a new perspective—the neighborhood life circle. This emerging urban decision-making and planning paradigm represents China’s attempt to address the “urban disease” arising from rapid urbanization recently, attracting global attention for its implementation of sustainability. This study aims to reveal the driving factors behind the livable neighborhood life circle amid rapid urbanization by conducting a systematic review of relevant empirical research within China’s context. We used Scopus and WoS as search databases, identifying and extracting a literature review of 67 publications from 2010 to 2025. The findings indicate that the driving factors of a livable neighborhood life circle are a structure constructed comprising social well-being, management and regulation, the built environment, and economic vitality, which are interconnected in multiple ways. This study has advanced discussions on the livable neighborhood life circle and expanded the existing knowledge and literature. It has also deepened insights into how sustainability concepts impact livable neighborhood life circles in China. The study offers insights into four aspects: the systematization of concepts and driving factors related to the neighborhood life circle in China, the development of assessment tools, the establishment of new planning paradigms, and the localization of implementation frameworks. Additionally, it further enriches the global application of the x-minute city and the neighborhood life circle. Full article

Mountain settlements represent culturally rich but environmentally fragile landscapes, shaped by enduring processes of ecological adaptation and human resilience. In western Sichuan, Jiarong Tibetan villages, with their distinctive integration of defensive stone towers and settlements, embody this coupling of culture and the environment. We hypothesize that settlement cores in these villages were shaped by natural environmental factors, with subsequent expansion reinforced by the cultural significance of towers. To test this, we applied a micro-scale spatial–environmental framework to four sample villages in Suopo Township, Danba County. High-resolution World Imagery (Esri, 0.5–1 m, 2022–2023) was classified via a Random Forest algorithm to generate detailed land-use maps, and a 100 × 100 m fishnet grid extracted topographic metrics (elevation, slope, aspect) and accessibility measures (distances to streams, roads, towers). Geographically weighted regression (GWR) was then used to examine how slope, elevation, aspect, proximity to water and roads, and tower distribution affect settlement patterns. The results show built-up density peaks on southeast-facing slopes of 15–30°, at altitudes of 2600–2800 m, and within 50–500 m of streams, co-locating with historic watchtower sites. Based on these findings, we propose four zoning strategies—a Core Protected Zone, a Construction And Development Zone, an Ecological Conservation Zone, and an Industry Development Zone—to balance preservation with growth. The resulting policy recommendations offer actionable guidance for sustaining traditional settlements in complex mountain environments. Full article

The rapid development of deep learning (DL) has demonstrated its modular contributions to artificial intelligence (AI) techniques, such as large language models (LLMs). DL variants have proliferated across domains such as feature extraction, normalization, lightweight architecture design, and module integration, yielding substantial advancements in these subfields. However, the absence of a unified review framework to contextualize DL’s modular evolutions within AI development complicates efforts to pinpoint future research directions. Existing review papers often focus on narrow technical aspects or lack systemic analysis of modular relationships, leaving gaps in our understanding how these innovations collectively drive AI progress. This work bridges this gap by providing a roadmap for researchers to navigate DL’s modular innovations, with a focus on balancing scalability and sustainability amid evolving AI paradigms. To address this, we systematically analyze extensive literature from databases including Web of Science, Scopus, arXiv, ACM Digital Library, IEEE Xplore, SpringerLink, Elsevier, etc., with the aim of (1) summarizing and updating recent developments in DL algorithms, with performance benchmarks on standard dataset; (2) identifying innovation trends in DL from a modular viewpoint; and (3) evaluating how these modular innovations contribute to broader advances in artificial intelligence, with particular attention to scalability and sustainability amid shifting AI paradigms. Full article

This study investigates the primary frequency regulation dynamic characteristics of industrial steam extraction turbine units under deep peak regulation conditions. A high-fidelity integrated dynamic model was established, incorporating the governor system, steam turbine with extraction modules, and interconnected pipeline dynamics. Through comparative simulations and experimental validation, the model demonstrates high accuracy in replicating real-unit responses to frequency disturbances. For the power grid system in this study, the frequency disturbance mainly comes from three aspects: first, the power imbalance formed by the random mutation of the load side and the intermittence of new energy power generation; second, transformation of the energy structure directly reduces the available frequency modulation resources; third, the system-equivalent inertia collapse effect caused by the integration of high permeability new energy; the rotational inertia provided by the traditional synchronous unit is significantly reduced. In the cogeneration unit and its control system in Guangxi involved in this article, key findings reveal that increased peak regulation depth (30~50% rated power) exacerbates nonlinear fluctuations. This is due to boiler combustion stability thresholds and steam pressure variations. Key parameters—dead band, power limit, and droop coefficient—have coupled effects on performance. Specifically, too much dead band (>0.10 Hz) reduces sensitivity; likewise, too high a power limit (>4.44%) leads to overshoot and slow recovery. The robustness of parameter configurations is further validated under source-load random-intermittent coupling disturbances, highlighting enhanced anti-interference capability. By constructing a coordinated control model of primary frequency modulation, the regulation strategy of boiler and steam turbine linkage is studied, and the optimization interval of frequency modulation dead zone, adjustment coefficient, and frequency modulation limit parameters are quantified. Based on the sensitivity theory, the dynamic influence mechanism of the key control parameters in the main module is analyzed, and the degree of influence of each parameter on the frequency modulation performance is clarified. This research provides theoretical guidance for optimizing frequency regulation strategies in coal-fired units integrated with renewable energy systems. Full article

Geum aleppicum Jacq. (yellow avens), a species traditionally used in folk medicine, remains understudied in the ethnopharmacological aspects. In this study, we comprehensively evaluated the phytochemical composition and biological activity of a hydroethanolic (50:50, v/v) extract from the aerial parts of G. aleppicum collected in Kazakhstan. Using the high-performance liquid chromatography coupled with electrospray ionization quadrupole time-of-flight tandem mass spectrometry (HPLC-ESI-QTOF-MS/MS), we identified 24 compounds, predominantly phenolic acids, flavonoids, tannins, and triterpenoids. The major compound was ellagic acid (2.28 mg/g dry extract) as revealed by the reverse phase high-performance liquid chromatography–diode array detector (RP-HPLC-DAD). The extract exhibited a high polyphenol content (131.45 mg GAE/g) and strong antioxidant activity in Ferric Reducing Antioxidant Power (FRAP) assay and 2,2-diphenyl-1-picrylhydrazyl (DPPH) radical scavenging assay (3.82 ± 0.07 mmol Fe²⁺/g and 106.61 ± 0.89 mg GAE/g, respectively). Antimicrobial assay of the extract revealed notable antifungal activity against Candida spp., especially against C. glabrata and C. tropicalis with minimum inhibitory concentration (MIC) of as low as 0.125 mg/mL, showing fungistatic effect. Although the extract inhibited the cytopathic effect induced by Human Herpesvirus 1 (HHV-1) in VERO cells, it did not significantly reduce viral replication. Moreover, among human cancer cell lines studied, the extract exerted moderate and selective cytotoxicity against A549 lung cancer cells (CC₅₀ = 75.51 µg/mL, SI = 9). These findings highlight G. aleppicum as a rich source of bioactive compounds, especially phenolics, supporting its potential for development of pharmaceutical and cosmetic applications. Full article

Most urban areas have Traffic Management Centers that rely partially on communication with 9-1-1 centers for incident detection. This level of awareness is often lacking for rural interstates spanning several 9-1-1 centers. This paper presents a novel approach to extending TMC visibility by automatically monitoring regional 9-1-1 dispatch channels using off-the-shelf hardware and open-source speech-to-text libraries. Our study presents a proof-of-concept study servicing 71 miles of rural I-65 in Indiana, successfully monitoring four county dispatch centers from a single location, and efficiently transcribing live audio within 60 s of broadcast. This work’s primary contribution is demonstrating the feasibility and practical value of automated incident detection systems for rural interstates. This technology is implementation-ready for extending the visibility of Traffic Management Centers in rural interstate segments. Further work is underway for developing scalable procedures for integrating multiple remote sites, extracting more diverse keyword sets, investigating optimal speech-to-text models, and assessing the technical aspects of the experimental procedures of this manuscript. Full article

By turning a waste stream into a clean energy source, green hydrogen generation from wastewater provides a dual solution to energy and environmental problems. This study presents a thorough bibliometric analysis of research trends in the field of green hydrogen generation from wastewater between 2010 and 2024. A total of 221 publications were extracted from Scopus database, and VOSviewer (1.6.20) was used as a visualization tool to identify influential authors, institutions, collaborations, and thematic focus areas. The analysis revealed a significant increase in research output, with a peak of 122 publications in 2024, with a total of 705 citations. China had the most contributions with 60 publications, followed by India (30) and South Korea (26), indicating substantial regional involvement in Asia. Keyword co-occurrence and coauthorship network mapping revealed 779 distinct keywords grouped around key themes like electrolysis, hydrogen evolution reactions, and wastewater treatment. Significantly, this work was supported by contributions from 115 publication venues, with the International Journal of Hydrogen Energy emerging as the most active and cited source (40 articles, 539 citations). The multidisciplinary aspect of the area was highlighted by keyword co-occurrence analysis, which identified recurring themes including electrolysis, wastewater treatment, and hydrogen evolution processes. Interestingly, the most-cited study garnered 131 citations and discussed the availability of unconventional water sources for electrolysis. Although there is growing interest in the field, it is still in its initial phases, indicating a need for additional research, particularly in developing countries. This work offers a basic overview for researchers and policymakers who are focused on promoting the sustainable generation of green hydrogen from wastewater. Full article

With the expansion of the scale of ultra-high-voltage transmission lines and the complexity of the corridor environment, the traditional manual inspection method faces serious challenges in terms of efficiency, cost, and safety. In this study, based on power laser inspection technology with a high-precision servo control system, a complete set of laser point cloud processing technology is proposed, covering three core aspects: transmission line extraction, scene recovery, and operation status monitoring. In transmission line extraction, combining the traditional clustering algorithm with the improved PointNet++ deep learning model, a classification accuracy of 92.3% is achieved in complex scenes; in scene recovery, 95.9% and 94.4% of the internal point retention rate of transmission lines and towers, respectively, and a vegetation denoising rate of 7.27% are achieved by RANSAC linear fitting and density filtering algorithms; in the condition monitoring segment, the risk detection of tree obstacles based on KD-Tree acceleration and the arc sag calculation of the hanging chain line model realize centimetre-level accuracy of hidden danger localisation and keep the arc sag error within 5%. Experiments show that this technology significantly improves the automation level and decision-making accuracy of transmission line inspection and provides effective support for intelligent operation and maintenance of the power grid. Full article