Audio-Text Multimodal Speech Recognition via Dual-Tower Architecture for Mandarin Air Traffic Control Communications

In air traffic control communications (ATCC), misunderstandings between pilots and controllers could result in fatal aviation accidents. Fortunately, advanced automatic speech recognition technology has emerged as a promising means of preventing miscommunications and enhancing aviation safety. However, most existing speech recognition methods merely incorporate external language models on the decoder side, leading to insufficient semantic alignment between speech and text modalities during the encoding phase. Furthermore, it is challenging to model acoustic context dependencies over long distances due to the longer speech sequences than text, especially for the extended ATCC data. To address these issues, we propose a speech-text multimodal dual-tower architecture for speech recognition. It employs cross-modal interactions to achieve close semantic alignment during the encoding stage and strengthen its capabilities in modeling auditory long-distance context dependencies. In addition, a two-stage training strategy is elaborately devised to derive semantics-aware acoustic representations effectively. The first stage focuses on pre-training the speech-text multimodal encoding module to enhance inter-modal semantic alignment and aural long-distance context dependencies. The second stage fine-tunes the entire network to bridge the input modality variation gap between the training and inference phases and boost generalization performance. Extensive experiments demonstrate the effectiveness of the proposed speech-text multimodal speech recognition method on the ATCC and AISHELL-1 datasets. It reduces the character error rate to 6.54% and 8.73%, respectively, and exhibits substantial performance gains of 28.76% and 23.82% compared with the best baseline model. The case studies indicate that the obtained semantics-aware acoustic representations aid in accurately recognizing terms with similar pronunciations but distinctive semantics. The research provides a novel modeling paradigm for semantics-aware speech recognition in air traffic control communications, which could contribute to the advancement of intelligent and efficient aviation safety management.

Swin-PAFF: A SAR Ship Detection Network with Contextual Cross-Information Fusion

Synthetic Aperture Radar(SAR)image target detection has widespread applications in both military and civil domains.However,SAR images pose challenges due to strong scattering,indistinct edge contours,multi-scale representation,sparsity,and severe background interference,which make the existing target detection methods in low accuracy.To address this issue,this paper proposes a multi-scale fusion framework(Swin-PAFF)for SAR target detection that utilizes the global context perception capability of the Transformer and the multi-layer feature fusion learning ability of the feature pyramid structure(FPN).Firstly,to tackle the issue of inadequate perceptual image context information in SAR target detection,we propose an end-to-end SAR target detection network with the Transformer structure as the backbone.Furthermore,we enhance the ability of the Swin Transformer to acquire contextual features and cross-information by incorporating a Swin-CC backbone network model that combines the Spatial Depthwise Pooling(SDP)module and the self-attentive mechanism.Finally,we design a cross-layer fusion neck module(PAFF)that better handles multi-scale variations and complex situations(such as sparsity,background interference,etc.).Our devised approach yields a noteworthy AP@0.5:0.95 performance of 91.3%when assessed on the HRSID dataset.The application of our proposed technique has resulted in a noteworthy advancement of 8%in the AP@0.5:0.95 scores on the HRSID dataset.

Electrochemical analysis and convection-enhanced mass transfer synergistic effect of MnO_x/Ti membrane electrode for alcohol oxidation

The different electrocatalytic reactors could be constructed for the electrocatalytic oxidation of 2,2,3,3-tetrafluoro-1-propanol(TFP) with two typical MnO_x/Ti electrodes, i.e.the electrocatalytic membrane reactor(ECMR) with the Ti membrane electrode and the electrocatalytic reactor(ECR) with the traditional Ti plate electrode.For the electro-oxidation of TFP, the conversion with membrane electrode(70.1%) in the ECMR was 3.3 and 1.7 times higher than that of the membrane electrode without permeate flow(40.8%) in the ECMR and the plate electrode(21.5%) in the ECR, respectively.Obviously, the pore structure of membrane and convection-enhanced mass transfer in the ECMR dramatically improved the catalytic activity towards the electro-oxidation of TFP.The specific surface area of porous electrode was 2.22 m^2·g^(-1).The surface area of plate electrode was 2.26 cm^2(1.13 cm^2× 2).In addition, the electrochemical results showed that the mass diffusion coefficient of the MnO_x/Ti membrane electrode(1.80 × 10^(-6) cm^2·s^(-1)) could be increased to 6.87 × 10^(-6) cm^2·s^(-1) at the certain flow rate with pump, confirming the lower resistance of mass transfer due to the convection-enhanced mass transfer during the operation of the ECMR.Hence, the porous structure and convection-enhanced mass transfer would improve the electrochemical property of the membrane electrode and the catalytic performance of the ECMR,which could give guideline for the design and application of the porous electrode and electrochemical reactor.

Characteristic wavelength selection of volatile organic compounds infrared spectra based on improved interval partial least squares

As important components of air pollutant,volatile organic compounds(VOCs)can cause great harm to environment and human body.The concentration change of VOCs should be focused on in real-time environment monitoring system.In order to solve the problem of wavelength redundancy in full spectrum partial least squares(PLS)modeling for VOCs concentration analysis,a new method based on improved interval PLS(iPLS)integrated with Monte-Carlo sampling,called iPLS-MC method,was proposed to select optimal characteristic wavelengths of VOCs spectra.This method uses iPLS modeling to preselect the characteristic wavebands of the spectra and generates random wavelength combinations from the selected wavebands by Monte-Carlo sampling.The wavelength combination with the best prediction result in regression model is selected as the characteristic wavelengths of the spectrum.Different wavelength selection methods were built,respectively,on Fourier transform infrared(FTIR)spectra of ethylene and ethanol gas at different concentrations obtained in the laboratory.When the interval number of iPLS model is set to 30 and the Monte-Carlo sampling runs 1000 times,the characteristic wavelengths selected by iPLS-MC method can reduce from 8916 to 10,which occupies only 0.22%of the full spectrum wavelengths.While the RMSECV and correlation coefficient(Rc)for ethylene are 0.2977 and 0.9999 ppm,and those for ethanol gas are 0.2977 ppm and 0.9999.The experimental results show that the iPLS-MC method can select the optimal characteristic wavelengths of VOCs FTIR spectra stably and effectively,and the prediction performance of the regression model can be significantly improved and simplified by using characteristic wavelengths.

Studies on Extracting Microcystin-Lr from Microcystis Aeruginosa by Water Bath

Different temperatures of water bath was used to extract the intracellular microcystin-LR(MC-LR) of Microcystis aeruginosa. Researching the extraction efficiency under the suitable temperature, so that it could find out the best temperature and time for extracting MC-LR from Microcystis aeruginosa cells. Five equal Microcystis aeruginosa was used to find out the best temperature, extracting at 60℃, 70℃, 80℃, 90℃ and 100℃ for 15 minutes, respectively. Results showed that the content of MC-LR extracted with the water under 100℃ was the highest. But meanwhile, the type and the content of impurities was the highest, too. In addition, another six equal Microcystis aeruginosa was extract with the water under 100℃ for 5min, 10 min, 15 min, 20 min, 25 min and 30 min respectively. It was proved that 20 minutes was enough for extracting MC-LR from Microcystis aeruginosa, no long time was needed.

ON THE UNIFORM STRONG APPROXIMATION OF MARCINKIEWICZ TYPE FOR MULTIVARIABLE CONTINUOUS FUNCTIONS

The rate of uniform strong approzimation of Marcinkiewicz type for multivariable continuous functions is obtained in this paper as follows：‖1/k＋1 ^k∑j=0 ｜Sj（f）-f｜^q‖≤C/k＋1^k∑j=0 Ej^q（f）,where Sj（f） denotes the square partial Fourier sum off and Ej（f） denotes the square best approximation of f by trigonometric polynomials of degree（j,j,…,j）,j=0,1,2.…

Study on the Concentration Inversion of NO &NO2 Gas from the Vehicle Exhaust Based on Weighted PLS

It becomes a key technology to measure the concentration of the vehicle exhaust components with the absorption spectra. But because of the overlap of gas absorption bands, how to separate the absorption information of each component gas from the mixed absorption spectra has become the key point to restrict the precision of the optical detection method. In this paper, the ex-perimental platform for the absorption spectrum of vehicle exhaust components has been established. Based on the ultraviolet absorption spectra measured with the platform of exhaust gas NO & NO2, the concentration regression model for the two components has been established with weighted partial least squares regression (WPLS). Finally the each spectral characteristic information of NO & NO2 gas has been separated and the concentration of each corresponding component has been reversed successfully.

Enhanced secondary organic aerosol formation during dust episodes by photochemical reactions in the winter in Wuhan

To investigate the effect of frequently occurring mineral dust on the formation of secondary organic aerosol(SOA),106 volatile organic compounds(VOCs),trace gas pollutants and chemical components of PM_(2.5) were measured continuously in January 2021 in Wuhan,Central China.The observation period was divided into two stages that included a haze period and a following dust period,based on the ratio of PM_(2.5) and PM_(10) concentrations.The average ratio of secondary organic carbon(SOC)to elemental carbon(EC)was 1.98 during the dust period,which was higher than that during the haze period(0.69).The contribution of SOA to PM_(2.5) also increased from 2.75% to 8.64%.The analysis of the relationships between the SOA and relative humidity(RH)and the odd oxygen(e.g.,O_(X)=O_(3)+NO_(2))levels suggested that photochemical reactions played a more important role in the enhancement of SOA production during the dust period than the aqueous-phase reactions.The heterogeneous photochemical production of OH radicals in the presence of metal oxides during the dust period was believed to be enhanced.Meanwhile,the ratios of trans-2-butene to cis-2-butene and m-/p-xylene to ethylbenzene(X/E)dropped significantly,confirming that stronger photochemical reactions occurred and SOA precursors formed efficiently.These results verified the laboratory findings that metal oxides in mineral dust could catalyse the oxidation of VOCs and induce higher SOA production.

秦岭国家植物园蝴蝶群落结构与多样性

为了查明秦岭国家植物园蝴蝶资源和多样性状况,本研究选取了5种生境,利用样线法进行了3年的蝴蝶多样性观测。结果表明:秦岭国家植物园蝴蝶共有5科76属134种;物种丰富度指数、物种多样性指数、均匀度指数和优势度指数分别为1.3295、3.4616、0.7068和0.7656;其中蛱蝶科属数及物种数最多(33属65种),凤蝶科最低(5属10种);蛱蝶科物种多样性指数及物种丰富度指数最高,分别为2.8525及1.3622;凤蝶科多样性指数最低(1.4936);粉蝶科物种丰富度指数最低(0.2790)。5种生境中,随着植被丰富度的增加,蝴蝶多样性指数上升。

Influence of Gas Temperature on Microstructure and Properties of Cold Spray 304SS Coating

In the present study, 304 stainless steel coatings were deposited on interstitial-free steel substrates by cold gas dynamic spray technology. The effect of gas temperature on microstructure, micro-hardness, cohesive strength, and electrochemical property of the coatings were investigated and compared. The results showed that increasing gas temperature had a great contribution to enhancing the bonding strength between the deposited particles and making the microstructure more density. Therefore, the porosity of the coatings decreased from 0%4-0.5% to 2%4-0.3%, and the tensile strength of the coatings increased from 564-4 MPa up to 734-3 MPa. In addition, the corrosion resistance of the coatings was also deeply influenced by process gas temperature. The corrosion kinetics of the coatings were affected by both of the plastic deformation of deposited particles and the porosity in the coatings.

A Decoupling Control Model on Perturbation Method for Twin-Roll Casting Magnesium Alloy Sheet

To better understand the twin-roll casting process,based on the analysis of the solidification phenomenon,the geometry shape of the molten metal pool,the continuity of metal and the balance of energy and momentum,five critical partial equations were established separately including the equations of pool level,solidification process,roll separating force,roll gap and casting speed.Meanwhile,to obtain a uniform sheet thickness and keep a constant roll separating force,a decoupling control model was built on the perturbation method to eliminate the interference of process parameters.The simulation results show that the control model is valuable to quickly and accurately determine the control parameters.Moreover,Mg alloy sheets with high quality were cast by applying this model.

Enhanced Fracture Toughness of Pressureless-sintered SiC Ceramics by Addition of Graphene

To enhance the fracture toughness of pressureless-sintered SiC ceramic, graphene was introduced as an additive. The effects of graphene contents on the fracture toughness, bending strength, micro-hardness, phase compositions, and microstructure evolutions of the SiC ceramics were investigated in detail by scanning electron microscopy, energy dispersive X-ray spectroscopy, and metallographic microscopy. The fracture toughness, bending strength and micro-hardness increased initially, and then decreased with the graphene content increasing from 0 to 5.0 wt%. The highest fracture toughness of 5.65 MPa m1/2 was obtained for sample with 1.0 wt% graphene sintered at 2130 ℃ for 1 h in At, which was about 22.6% higher than that of SiC sample without graphene. In addition, the highest bending strength and microhardness of 434.14 MPa and 29.21 GPa corresponded to the SiC samples with graphene content of 0.5 wt% and 2.0 wt%, respectively.

Embryonic expression and evolutionary analysis of the amphioxus Dickkopf and Kremen family genes

The secreted Wnt signaling inhibitor Dickkopfl （Dkkl） plays key role in vertebrate head induction. Its receptor Kremen synergizes with Dkkl in Wnt inhibition. Here we have carried out expression and functional studies of the Dkk and Kremen genes in amphioxus （Branchiostoma belcheri）. During embryonic and larval development, BbDkkl/2/4 is expressed in the posterior mesoendoderm, anterior somatic mesoderm and the pharyngeal regions. Its expression becomes restricted to the pharyngeal region on the left side at larval stages. In 45 h larvae, BbDkkl/2/4 is expressed specifically in the cerebral vesicle. BbDkk3 was only detected at larval stages in the mid-intestine region. Seven Kremen related genes were identified in the genome of the Florida amphioxus （Branchiostoma floridae）, clustered in 4 scaffolds, and are designated Kremenl-4 and Kremen-like 1-3, respectively. In B. belcheri, Kremenl is strongly expressed in the mesoendoderm during early development and Kremen3 is expressed asymmetrically in spots in the larval pharyngeal region. In luciferase reporter assays, BbDkkl/2/4 can strongly inhibit Wnt signaling, while BbDkk3, BbKremenl and BbKremen3 can not. No co-operative effect was observed between amphioxus Dkkl/2/4 and Kremens, suggesting that the interaction between Dkk and Kremen likely originated later during evolution.

Exchange coupling-induced uniaxial anisotropy in La_(0.7)Sr_(0.3)MnO_3 thin films

Establishing a deeper understanding of the anisotropy in manganites is useful for tailoring their magnetic properties for device applications. Here we showed that ferromagnetic La_(0.7)Sr_(0.3)MnO_3(LSMO) thin films,epitaxially grown on SrTiO_3(STO) substrates, exhibited unexpected double-shifted magnetization curves originating from the competition between common biaxial and emergent uniaxial anisotropies. This emergent uniaxial anisotropy could be induced by exchange coupling between the ferromagnetic LSMO and an antiferromagnetic LSMO dead layer at the LSMO/STO interface, which could be manipulated by the degree of oxygen deficiency.

FEM ANALYSIS FOR INFLUENCES OF A FAULT ON COUPLED T-H-M-M PROCESS IN DUAL-POROSITY ROCK MASS

For the case in which a large geological structure like fault existing within the surrounding rock mass in the near field of a repository for high-level radioactive nuclear waste, one kind of coupled thermo-hydro-mechanical-migratory model of dual-porosity medium for saturated- unsaturated ubiquitous-joint rock mass was established. In the present model, the seepage field and the concentration field are double, but the stress field and the temperature field are single, and the influences of sets, spaces, angles, continuity ratios, stiffness of fractures on the constitutive relationship of the medium can be considered. At the same time, a two-dimensional program of finite element method was developed. Taking a hypothetical nuclear waste repository located at a rock mass being unsaturated dual-porosity medium as a calculation example, the FEM analysis for thermo-hydro-mechanical-migratory coupling were carried out under the condition of radioac- tive nuclide leaking for the cases with and without a fault, and the temperatures, pore pressures, flow velocities, nuclide concentrations and principal stresses in the rock mass were investigated. The results show that the fracture water in the fault flows is basically along the fault direction, and its flow velocity is almost three orders of magnitude higher than that of fracture water in rock mass; the nuclide concentration in the fault is also much higher than that without fault, and the nuclides move along the fault faster; moreover, the fault has obvious influences on the pore pressures and the principal stresses in the rock mass.

aunalysis of the Genome Sequence of the Medicinal Plant Salvia miltiorrhiza

Dear Editor Salvia miltiorrhiza Bunge （Danshen） is a medicinal plant of the Lamiaceae family, and its dried roots have long been used in traditional Chinese medicine with hydrophilic phenolic acids and tanshinones as pharmaceutically active components （Zhang et al., 2014; Xu et al., 2016）. The first step of tanshinone biosynthesis is bicyclization of the general diterpene precursor （E,E,E）-geranylgeranyl diphosphate （GGPP） to copalyl diphosphate （CPP） by CPP synthases （CPSs）, which is followed by a cyclization or rearrangement reaction catalyzed by kaurene synthase-like enzymes （KSL）. The resulting intermediate is usually an olefin, which requires the insertion of oxygen by cytochrome P450 mono-oxygenases （CYPs） for the final production of diterpenoids （Zi et al., 2014）. While the CPS, KSL, and several early acting CYPs （CYP76AH1, CYP76AH3, and CYP76AK1） for tanshinone biosynthesis have been identified in S. miltiorrhiza （Gao et al., 2009; Guo et al., 2013, 2016; Zi and Peters, 2013）, the majority of the overall biosynthetic pathway, as well as the relevant regulatory factors associated with tanshinone production, remains elusive （Figure 1B）.

Panorama of boron nitride nanostructures via lamp ablation

A diverse range of remarkable boron n itride (BN) nano structures subsumi ng nano-homs, nano-rods, nano-platelets, and clusters of hollow nano spheres (nano-onions, arguably of greatest applied and fun damental in terest) have bee n produced exclusively from crystalline BN precursor powder via lamp ablation. The procedure is safe, devoid of toxic reagents, simple, rapid and scalable—generating some genres of nan oparticles that had previously proved elusive. Product structure and composition were unambiguously assessed by high-resoluti on transmission electron microscopy, energy-dispersive X-ray spectroscopy and electron energy loss spectroscopy.

Dynamic image acquisition and particle recognition of laser-induced exit surface particle ejection in fused silica

Particle ejection is an important process during laser-induced exit surface damage in fused silica. Huge quantities of ejected particles, large ejection velocity, and long ejection duration make this phenomenon difficult to be directly observed. An in situ two-frame shadowgraphy system combined with a digital particle recognition algorithm was employed to capture the transient ejecting images and obtain the particle parameters. The experimental system is based on the principle of polarization splitting and can capture two images at each damage event. By combining multiple similar damage events at different time delays, the timeline of ejecting evolution can be obtained. Particle recognition is achieved by an adaptively regularized kernel-based fuzzy C-means algorithm based on a grey wolf optimizer. This algorithm overcomes the shortcoming of the adaptively regularized kernel-based fuzzy C-means algorithm easily falling into the local optimum and can resist strong image noises, including diffraction pattern, laser speckle, and motion artifact. This system is able to capture particles ejected after 600 ns with a time resolution of 6 ns and spatial resolution better than 5 μm under the particle recognition accuracy of 100%.

Characteristics of optical emission during laser-induced damage events of fused silica

Spontaneous optical emission properties of laser-produced plasma during laser damage events at input and exit surfaces of fused silica were retrieved and compared.We show that plasma at the input surface is much larger in size and exhibits significantly higher electron number density and excitation temperature,even when smaller laser energy was used.This effect was attributed to the stronger laser–plasma coupling at the input surface.In addition,a strong continuum background containing three peaks at 1.3eV,1.9eV,and 2.2eV was observed at the exit surface,and possible origins for this effect are also discussed.