Single Domain Research Articles

Comparative study in magnetic properties of FeCo alloy nanowires and a first-order reversal curve analysis

FeCo alloy nanowires have potential applications in data storage, magnetic sensors and spintronic devices. In this paper, FeCo alloy nanowires with varying diameters were fabricated within an anodic aluminum oxide (AAO) template using electrodeposition method at a constant potential. Characterization of the morphology and structure revealed that the polycrystalline body-center-cubic(bcc) FeCo alloy nanowires were obtained. The influence of diameter of FeCo alloy nanowires on magnetic interaction, domain state and magnetic mechanism was investigated through a first-order reversal curve (FORC) analysis. The findings showed that with an increase in the diameter of FeCo alloy nanowires from 20 nm to 60 nm and 100 nm, the interaction force between the FeCo alloy nanowires increased, the domain state of FeCo alloy nanowires changed from Single Domain (SD) to Multi-Domain (MD), the magnetic reversal mode changed from coherent reversal mode to curling reversal mode.

Single domain antibody: Development and application in biotechnology and biopharma.

Heavy-chain antibodies (HCAbs) are a unique type of antibodies devoid of light chains, and comprised of two heavy chains-only that recognize their cognate antigen by virtue of a single variable domain also referred to as VHH, single domain antibody (sdAb), or nanobody (Nb). These functional HCAbs, serendipitous discovered about three decades ago, are exclusively found in camelids, comprising dromedaries, camels, llamas, and vicugnas. Nanobodies have become an essential tool in biomedical research and medicine, both in diagnostics and therapeutics due to their beneficial properties: small size, high stability, strong antigen-binding affinity, low immunogenicity, low production cost, and straightforward engineering into more potent affinity reagents. The occurrence of HCAbs in camelids remains intriguing. It is believed to be an evolutionary adaptation, equipping camelids with a robust adaptive immune defense suitable to respond to the pressure from a pathogenic invasion necessitating a more profound antigen recognition and neutralization. This evolutionary innovation led to a simplified HCAb structure, possibly supported by genetic mutations and drift, allowing adaptive mutation and diversification in the heavy chain variable gene and constant gene regions. Beyond understanding their origins, the application of nanobodies has significantly advanced over the past 30 years. Alongside expanding laboratory research, there has been a rapid increase in patent application for nanobodies. The introduction of commercial nanobody drugs such as Cablivi, Nanozora, Envafolimab, and Carvykti has boosted confidence among in their potential. This review explores the evolutionary history of HCAbs, their ontogeny, and applications in biotechnology and pharmaceuticals, focusing on approved and ongoing medical research pipelines.

Skip-patching spatial–temporal discrepancy-based anomaly detection on multivariate time series

Anomaly detection in the Industrial Internet of Things (IIoT) is a challenging task that relies heavily on the efficient learning of multivariate time series representations. We introduce Skip-patching and Spatial–Temporal discrepancy mechanisms to improve the efficiency of detecting anomalies. Traditional feature extraction is hindered by redundant information in limited datasets. The situation is that feature generation from stable operational processes results in low-quality representations. To address this challenge, we propose the Skip-Patching mechanism. This approach involves selectively extracting features from partial data patches, prompting the model to learn more meaningful knowledge through self-supervised learning. It also effectively doubles the training sample size by creating independent sub-groups of patches. Despite the complex spatial and temporal relationships in IIoT systems, existing methods mainly extracted features from a single domain, either temporal or spatial (sensor-wise), or simply cascaded two features, i.e., one after one, which limited anomaly detection capabilities. To address this, we introduce the Spatial–Temporal Association Discrepancy component, which leverages discrepancies between spatial and temporal features to enhance latent representation learning. Our Skip-Patching Spatial–Temporal Anomaly Detection (SSAD) framework combines these two components to provide a more diverse and comprehensive learning process. Tested across four multivariate time series anomaly detection benchmarks, SSAD demonstrates superior performance, confirming the efficacy of combining Skip-patching and Spatial–Temporal features to enhance anomaly detection in IIoT systems.

Stress induced martensitic transformation in NiTi at elevated temperatures: Martensite variant microstructures, recoverable strains and plastic strains

To shed light on the origin of the loss of functional properties of NiTi with temperature increasing above 100 °C, we have investigated stress induced martensitic transformations in nanocrystalline NiTi shape memory wire by thermomechanical tensile testing supplemented with post-mortem reconstruction of martensite variant microstructures in grains by nanoscale orientation mapping in TEM. The stress induced martensitic transformation generating recoverable transformation strain as well as plastic strain is not completed at the end of the upper stress plateau. The higher is the test temperature, the larger is the volume fraction of retained austenite as well as the plastic strain. The martensite variant microstructures in NiTi wire deformed up to the end of the stress plateau at 120 °C contain partially detwinned single domains of (001) compound twin laminate filling entire grains. It is proposed that the stress induced martensitic transformation proceeds via habit plane interface between austenite and second order laminate of (001) compound twins and that the martensite promptly reorients and deforms plastically by dislocation glide in the [100](001) slip system. When the wire is loaded further beyond the end of the stress plateau, the stress induced martensitic transformation continues and the martensite deforms plastically. It is concluded that the observed gradual loss of superelastic functionality of NiTi with increasing temperature does not originate from the plastic deformation of austenite, as widely assumed in the literature, but that it derives from the loss of resistance of the stress induced martensite to the plastic deformation under increasing stress.

Y-switch: a spring-loaded synthetic gene switch for robust DNA/RNA signal amplification and detection.

Nucleic acid tests (NATs) are essential for biomedical diagnostics. Traditional NATs, often complex and expensive, have prompted the exploration of toehold-mediated strand displacement (TMSD) circuits as an economical alternative. However, the wide application of TMSD-based reactions is limited by 'leakage'-the spurious activation of the reaction leading to high background signals and false positives. Here, we introduce the Y-Switch, a new TMSD cascade design that recognizes a custom nucleic acid input and generates an amplified output. The Y-Switch is based on a pair of thermodynamically spring-loaded DNA modules. The binding of a predefined nucleic acid target triggers an intermolecular reaction that activates a T7 promoter, leading to the perpetual transcription of a fluorescent aptamer that can be detected by a smartphone camera. The system is designed to permit the selective depletion of leakage byproducts to achieve high sensitivity and zero-background signal in the absence of the correct trigger. Using Zika virus (ZIKV)- and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)-derived nucleic acid sequences, we show that the assay generates a reliable target-specific readout. Y-Switches detect native RNA under isothermal conditions without reverse transcription or pre-amplification, with a detection threshold as low as ∼200 attomole. The modularity of the assay allows easy re-programming for the detection of other targets by exchanging a single sequence domain. This work provides a low-complexity and high-fidelity synthetic biology tool for point-of-care diagnostics and for the construction of more complex biomolecular computations.

Pam16 and Pam18 were repurposed during Trypanosoma brucei evolution to regulate the replication of mitochondrial DNA.

Protein import and genome replication are essential processes for mitochondrial biogenesis and propagation. The J-domain proteins Pam16 and Pam18 regulate the presequence translocase of the mitochondrial inner membrane. In the protozoan Trypanosoma brucei, their counterparts are TbPam16 and TbPam18, which are essential for the procyclic form (PCF) of the parasite, though not involved in mitochondrial protein import. Here, we show that during evolution, the 2 proteins have been repurposed to regulate the replication of maxicircles within the intricate kDNA network, the most complex mitochondrial genome known. TbPam18 and TbPam16 have inactive J-domains suggesting a function independent of heat shock proteins. However, their single transmembrane domain is essential for function. Pulldown of TbPam16 identifies a putative client protein, termed MaRF11, the depletion of which causes the selective loss of maxicircles, akin to the effects observed for TbPam18 and TbPam16. Moreover, depletion of the mitochondrial proteasome results in increased levels of MaRF11. Thus, we have discovered a protein complex comprising TbPam18, TbPam16, and MaRF11, that controls maxicircle replication. We propose a working model in which the matrix protein MaRF11 functions downstream of the 2 integral inner membrane proteins TbPam18 and TbPam16. Moreover, we suggest that the levels of MaRF11 are controlled by the mitochondrial proteasome.

1H, 13C and 15N resonance assignments of a shark variable new antigen receptor against hyaluronan synthase.

Single domain antibody (sdAb) is only composed of a variable domain of the heavy-chain-only antibody, which is devoid of light chain and naturally occurring in camelids and cartilaginous fishes. Variable New Antigen Receptor (VNAR), a type of single domain antibody present in cartilaginous fishes such as sharks, is the smallest functional antigen-binding fragment found in nature. The unique features, including flexible paratope, high solubility and outstanding stability make VNAR a promising prospect in antibody drug development and structural biology research. However, VNAR's research has lagged behind camelid-derived sdAb, especially in the field of structural research. Here we report the 1H,15N,13C resonance assignments of a VNAR derived from the immune library of Chiloscyllium plagiosum, termed B2-3, which recognizes the hyaluronan synthase. Analysis of the backbone chemical shifts demonstrates that the secondary structure of VNAR is predominately composed of β-sheets corresponding to around 40% of the B2-3 backbone. The Cβ chemical shift values of cysteine residues, combined with mass spectrometry data, clearly shows that B2-3 contains two pairs of disulfide bonds, which is import for protein stability. The assignments will be essential for determining the high resolution solution structure of B2-3 by NMR spectroscopy.

An improved cross-domain sequential recommendation model based on intra-domain and inter-domain contrastive learning

Cross-domain recommendation aims to integrate data from multiple domains and introduce information from source domains, thereby achieving good recommendations on the target domain. Recently, contrastive learning has been introduced into the cross-domain recommendations and has obtained some better results. However, most cross-domain recommendation algorithms based on contrastive learning suffer from the bias problem. In addition, the correlation between the user’s single-domain and cross-domain preferences is not considered. To address these problems, a new recommendation model is proposed for cross-domain scenarios based on intra-domain and inter-domain contrastive learning, which aims to obtain unbiased user preferences in cross-domain scenarios and improve the recommendation performance of both domains. Firstly, a network enhancement module is proposed to capture users’ complete preference by applying a graphical convolution and attentional aggregator. This module can reduce the limitations of only considering user preferences in a single domain. Then, a cross-domain infomax objective with noise contrast is presented to ensure that users’ single-domain and cross-domain preferences are correlated closely in sequential interactions. Finally, a joint training strategy is designed to improve the recommendation performances of two domains, which can achieve unbiased cross-domain recommendation results. At last, extensive experiments are conducted on two real-world cross-domain scenarios. The experimental results show that the proposed model in this paper achieves the best recommendation results in comparison with existing models.

Rolling Bearing Fault Diagnosis in Agricultural Machinery Based on Multi-Source Locally Adaptive Graph Convolution

Maintaining agricultural machinery is crucial for efficient mechanized farming. Specifically, diagnosing faults in rolling bearings, which are essential rotating components, is of significant importance. Domain-adaptive technology often addresses the challenge of limited labeled data from a single source domain. However, information transfer can sometimes fall short in providing adequate relevant details for supporting target diagnosis tasks, leading to poor recognition performance. This paper introduces a novel fault diagnosis model based on a multi-source locally adaptive graph convolution network to diagnose rolling bearing faults in agricultural machinery. The model initially employs an overlapping sampling method to enhance sample data. Recognizing that two-dimensional time–frequency signals possess richer spatial characteristics in neural networks, wavelet transform is used to convert time series samples into time–frequency graph samples before feeding them into the feature network. This approach constructs a sample data pair from both source and target domains. Furthermore, a feature extraction network is developed by integrating the strengths of deep residual networks and graph convolutional networks, enabling the model to better learn invariant features across domains. The locally adaptive method aids the model in more effectively aligning features from the source and target domains. The model incorporates a Softmax layer as the bearing state classifier, which is set up after the graph convolutional network layer, and outputs bearing state recognition results upon reaching a set number of iterations. The proposed method’s effectiveness was validated using a bearing dataset from Jiangnan University. For three different groups of bearing fault diagnosis tasks under varying working conditions, the proposed method achieved recognition accuracies above 99%, with an improvement of 0.30%-4.33% compared to single-source domain diagnosis models. Comparative results indicate that the proposed method can effectively identify bearing states even without target domain labels, showcasing its practical engineering application value.

Long-term outcomes and adaptive behavior in adult patients with Lennox-Gastaut syndrome.

Lennox-Gastaut syndrome (LGS) is a severe form of epilepsy characterized by difficult-to-control seizures and cognitive dysfunction. Previous studies mainly focused on pediatric populations, and little is known about the long-term cognitive outcome in adult patients with LGS. The objective of this study was to investigate the long-term functional and adaptive behavior in adult patients with LGS. This cross-sectional study enrolled adult patients diagnosed with LGS according to the recently published International League Against Epilepsy (ILAE) diagnostic criteria. The adaptive behavior of participants was assessed using the Vineland Adaptive Behavior Scales, Survey Interview, Second Edition (VABS-II). Demographic, clinical, electroencephalography (EEG), and antiseizure medication (ASM) data were also collected at different timepoints, to investigate their association with VABS-II scores. The study included 38 adult patients with LGS. A low score on the Adaptive Behavior Composite Scale was found in all patients. When considering single VABS-II domains, particularly low scores were found in daily living skills and socialization, whereas slightly higher performances were observed in communication. An earlier age at LGS diagnosis was identified as the most significant predictor of worse adaptive outcomes in adult life. At the time of study evaluation, high seizure frequency, higher EEG background slowing, and multifocal EEG epileptiform abnormalities were significantly associated with lower VABS-II raw scores. Furthermore, in an exploratory correlation analysis with ASM regimen at the study visit, treatment with cannabidiol was associated with higher adaptive behavior scores, whereas benzodiazepine intake correlated with lower scores. This study provides relevant insights into the long-term challenges faced by adults with Lennox-Gastaut syndrome (LGS), highlighting significant impairments in adaptive behavior as well as the associated clinical and electroencephalography features. Additionally, this study provides a more specific neuropsychological profile in adults with LGS and underscores the importance of comprehensive care approaches that go beyond seizure control in this population. This study examined adults with Lennox-Gastaut syndrome (LGS), a severe type of epilepsy, to understand their long-term abilities to perform daily tasks and adapt socially. We found that these adults have significant difficulties with daily living and social skills, although not all areas were equally affected. They performed somewhat better in communication, particularly in understanding others (receptive communication). Importantly, the younger the age at which LGS was diagnosed, the worse their outcomes were as adults. This study highlights the need for research and treatment approaches that focus not only on controlling seizures but also on improving daily life skills.

Comprehensive characterization and expression profiling of BBX gene family in soybean in response to UV-B stress

The BBX gene family encodes transcription factors that regulate several biological processes, such as plant growth, development, and stress response. They have been identified and thoroughly studied in numerous plant species. The functional role of the BBX gene family against UV-B stress in soybean has not been well documented. Herein, this study focused on a comprehensive investigation of the GmBBX genes in soybean and their expression patterns in response to UV-B exposure. A total of 184 BBX genes were identified and analyzed to determine the conserved domain. Single and double B-box and CCT domains were found across these proteins, indicating potential functional diversity. The protein physicochemical properties showcased diverse characteristics, highlighting molecular size, stability, and hydrophobicity. Phylogenetic analysis revealed five distinct clades, showcasing a non-uniform distribution of GmBBX genes. Chromosomal mapping showed a non-uniform distribution of GmBBX genes across the 20 chromosomes of soybean. The highest number of genes (26) were found on chromosome 13, whereas chromosomes 5 and 15 displayed the lowest number of genes (3 genes each). Gene structure analysis revealed variations in exon-intron patterns, while motif composition analysis unveiled conserved motifs among GmBBX proteins. Gene duplication indicated that the expansion of the GmBBX genes was linked with tandem and dispersed duplications. Gene ontology analysis revealed enrichment of terms associated with light stimulus, response to abiotic stress, and molecular functions such as zinc ion binding. The promoter region reveals cis-acting elements related to hormones, light, and stress responses. The GmBBX genes exhibited differential expression in response to UV-B stress, suggesting their possible role in plant defense against UV-B stress. This study will provide a solid foundation for further analysis of the molecular mechanisms of GmBBX genes in soybean against UV-B stress.

Well-Defined Homopolymer Nanoparticles with Uniaxial Molecular Orientation by Synchronized Polymerization and Self-Assembly.

Synthesizing anisotropic polymeric nanoparticles (NPs) with well-defined shapes, dimensions, and molecular orientations is a very challenging task. Herein, we report the synthesis of surprisingly highly uniform shape-anisotropic polymer NPs with uniaxial internal molecular orientation. Keys to our method are synchronized polymerization and self-assembly (SPSA), which can even be realized by regular dispersion polymerization. This is demonstrated using a monomer containing a rigid 4-nitroazobenzene (NAB) side group. The short nucleation period, the completion of microphase separation before molecular motion is frozen, and sufficient low particle/solvent interfacial tension are shown to be the origins of the highly uniform dimensions, single liquid crystal domains, and well-defined anisotropic shape of particles. The liquid crystallization ability of the polymers, control of molecular weight distribution, and the polymerization kinetics are identified as three key factors controlling the NP formation. The uniformity of these NPs facilitates their SA formation into colloidal crystals. The particles exhibit optically anisotropic properties depending on orientations and, in particular, show intriguing photoswitchable LC-glass (order-disorder) transition, which can be used for the detection of ultraviolet (UV) light and allows the fabrication of photoreversible colloidal films.

Nanoscale Imaging and Measurements of Grain Boundary Thermal Resistance in Ceramics with Scanning Thermal Wave Microscopy.

Material thermal conductivity is a key factor in various applications, from thermal management to energy harvesting. With microstructure engineering being a widely used method for customizing material properties, including thermal properties, understanding and controlling the role of extended phonon-scattering defects, like grain boundaries, is crucial for efficient material design. However, systematic studies are still lacking primarily due to limited tools. In this study, we demonstrate an approach for measuring grain boundary thermal resistance by probing the propagation of thermal waves across grain boundaries with a temperature-sensitive scanning probe. The method, implemented with a spatial resolution of about 100 nm on finely grained Nb-substituted SrTiO3 ceramics, achieves a detectability of about 2 × 10-8 K m2 W-1, suitable for chalcogenide-based thermoelectrics. The measurements indicated that the thermal resistance of the majority of grain boundaries in the STiO3 ceramics is below this value. While there are challenges in improving sensitivity, considering spatial resolution and the amount of material involved in the detection, the sensitivity of the scanning probe method is comparable to that of optical thermoreflectance techniques, and the method opens up an avenue to characterize thermal resistance at the level of single grain boundaries and domain walls in a spectrum of microstructured materials.

S. aureus Eap is a polyvalent inhibitor of neutrophil serine proteases

Staphylococcus aureus expresses three high-affinity neutrophil serine protease (NSP) inhibitors known as the extracellular adherence protein domain (EAPs) proteins. Whereas EapH1 and EapH2 are comprised of a single EAP domain, the modular extracellular adherence protein (Eap) from S. aureus strain Mu50 consists of four EAP domains. We recently reported that EapH2 can simultaneously bind and inhibit cathepsin-G (CG) and neutrophil elastase (NE), which are the two most abundant NSPs. This unusual property of EapH2 arises from independent CG and NE-binding sites that lie on opposing faces of its EAP domain. Here we used X-ray crystallography and enzyme assays to show that all four individual domains of Eap (i.e. Eap1, Eap2, Eap3, and Eap4) exhibit an EapH2-like ability to form ternary complexes with CG and NE that inhibit both enzymes simultaneously. We found that Eap1, Eap2, and Eap3 have similar functional profiles insofar as NSP inhibition is concerned, but that Eap4 displays an unexpected ability to inhibit two NE enzymes simultaneously. Using X-ray crystallography, we determined that this second NE-binding site in Eap4 arises through the same region of its EAP domain that also comprises its CG-binding site. Interestingly, small angle X-ray scattering data showed that stable tail-to-tail dimers of the NE/Eap4/NE ternary complex exist in solution. This arrangement is compatible with NSP-binding at all available sites in a two-domain fragment of Eap. Together, our work implies that Eap is a polyvalent inhibitor of NSPs. It also raises the possibility that higher-order structures of NSP-bound Eap may have unique functional properties.

Surface structure of MOVPE-prepared As-modified Si(100) substrates

In the pursuit of high-efficiency tandem devices for solar energy conversion based on III-V-semiconductors, low-defect III-V nucleation on Si(100) substrates is essential. Here, hydrogen and arsenic are key ingredients in all growth processes with respect to industrially scalable metalorganic vapor phase epitaxy. Our study provides insight into Si(100) surface preparation for the initial stage of III-V nucleation. The samples investigated, prepared on substrates with different offcut angles, show single domain surfaces consisting of rows of preferentially buckled dimers. Low energy electron diffraction and reflection anisotropy spectroscopy confirm well-defined (1 × 2)/(2 × 1) majority domains. Fourier-transform infrared spectroscopy revealed hydrogen bonding to the surface dimers, while no impurities were found by XPS. Density functional theory calculations support the experimental results and reveal a novel surface motif of H-passivated Si-As mixed dimers.

The process using a synthetic library that generates multiple diverse human single domain antibodies.

Single domain antibodies (sdAbs) possess unique characteristics that make them highly effective for developing complex therapeutics. Our process uses a fully synthetic phage display library to generate single domain antibodies that can bind to disease relevant antigen conformations. A human IGHV3 family scaffold makes up the phage display libraries, and these VHO libraries are applied to diverse phage biopannings against target antigens. After NGS processing, unique VHOs undergo automated cloning into expression constructs followed by transfections and purifications. Binding assays were used to determine VHO binding behaviors to the target proteins. Additional VHO interactions are measured against endogenous targets on cells by way of flow cytometry, cell internalization, and activation assays. We show that a fully synthetic phage display library can generate VHOs that bind to disease relevant antigen conformations. The diverse biopanning methods and processing of next-generation sequencing generated many VHO paratopes. These different VHO sequences can be expressed as Fc fusion proteins. Various screening assays resulted in VHOs representing different epitopes or activities. During the hit evaluation, we demonstrate how screening can identify distinct VHO activities that have been used to generate differentiated drug molecules in various bispecific and multispecific antibody formats. We demonstrate how screening can identify distinct VHO activities that have been used to generate differentiated drug molecules in various bispecific and multispecific antibody formats.

A new generation of nanobody research tools using improved mass spectrometry-based discovery methods

Single domain antibodies (“nanobodies”) derived from the variable region of camelid heavy-chain only antibody variants have proven to be widely useful tools for research, as well as therapeutic and diagnostic applications. In addition to traditional display techniques, methods to generate nanobodies using direct detection by mass spectrometry and DNA sequencing have been highly effective. However, certain technical challenges have limited widespread application. We have optimized a new pipeline for this approach that greatly improves screening sensitivity, depth of antibody coverage, antigen compatibility, and overall hit rate and affinity. We have applied this improved methodology to generate significantly higher affinity nanobody repertoires against widely used targets in biological research – i.e., GFP, tdTomato, GST, and mouse, rabbit, and goat IgG. We have characterized these reagents in affinity isolations and tissue immunofluorescence microscopy, identifying those that are optimal for these particularly demanding applications, and engineering dimeric constructs for ultra-high affinity. This study thus provides new nanobody tools directly applicable to a wide variety of research problems, and improved techniques enabling future nanobody development against diverse targets.

Vortex Magnetic Domain State Behavior in the Day Plot

AbstractThe ability of rocks to hold a reliable record of the ancient geomagnetic field depends on the structure and stability of magnetic domain‐states contained within constituent particles. In paleomagnetic studies, the Day plot is an easily constructed graph of magnetic hysteresis parameters that is frequently used to estimate the likely magnetic recording stability of samples. Often samples plot in the region of the Day plot attributed to so‐called pseudo‐single‐domain particles with little understanding of the implications for domain‐states or recording fidelity. Here we use micromagnetic models to explore the hysteresis parameters of magnetite particles with idealized prolate and oblate truncated‐octahedral geometries containing single domain (SD), single‐vortex and occasionally multi‐vortex states. We show that these domain states exhibit a well‐defined trend in the Day plot that extends from the SD region well into the multi‐domain region, all of which are likely to be stable remanence carriers. We suggest that although the interpretation of the Day plot and its variants might be subject to ambiguities, if the magnetic mineralogy is known, it can still provide some useful insights about paleomagnetic specimens' dominant domain state, average particle sizes and, consequently, their paleomagnetic stability.

Social Media Virality: Reaching the Tipping Point

Social media virality has become a key factor in determining how far to extend social media marketing and digital promotion campaigns. The point at which content transitions to viral spread underpins promotion-level decisions. We employ a logistic equation to identify tipping point decisions for viral content online promotions in given contexts. We consider case studies indicative of viral content on social media following a traditional S-curve when examined in a single domain. Our study contributes to understanding the decision mechanics of reaching a tipping point in a manner that reflects the concerns of digital marketers and online marketing strategies generally. We also see our approach to determining this point as furthering social media research in an area of growing importance and assessing how viral content can be managed.

Multi-scale hierarchical model for long-term time series forecasting

Long-term time series forecasting (LTSF) has become an urgent requirement in many applications, such as wind power supply planning. This is a highly challenging task because it requires considering both the complex frequency-domain and time-domain information in long-term time series simultaneously. However, existing work only considers potential patterns in a single domain (e.g., time or frequency domain), whereas a large amount of time-frequency domain information exists in real-world LTSFs. In this paper, we propose a multi-scale hierarchical network (MHNet) based on time-frequency decomposition to solve the above problem. MHNet first introduces a multi-scale hierarchical representation, extracting and learning features of time series in the time domain, and gradually builds up a global understanding and representation of the time series at different time scales, enabling the model to process time series over lengthy periods of time with lower computational complexity. Then, the robustness to noise is enhanced by employing a transformer that leverages frequency-enhanced decomposition to model global dependencies and integrates attention mechanisms in the frequency domain. Meanwhile, forecasting accuracy is further improved by designing a periodic trend decomposition module for multiple decompositions to reduce input-output fluctuations. Experiments on five real benchmark datasets show that the forecasting accuracy and computational efficiency of MHNet outperform state-of-the-art methods.