Dense Network Research Articles

Data-driven urban configuration optimization: An XGBoost-based approach for mitigating flood susceptibility and enhancing economic contribution

The indiscriminate evolution of urban configurations aggravates flood vulnerabilities, threatening sustainable urban expansion. Present methodologies fall short in supplying urban planners with flood mitigative strategies centered on urban configuration facets. Leveraging the power of the XGBoost algorithm, this study posits an advanced optimization schema, adroitly balancing the dual objectives of mitigating urban flooding and enhancing economic growth, with minimal disruption to established urban layouts. Shenzhen serves as the investigative ground, where the model displays exceptional accuracy, resilience, and interpretability in predicting Pluvial Flooding Susceptibility (PFS) and Economic Contribution (EC). Model interpretation divulges the profound influence of three-dimensional urban configuration elements, primarily the Building Congestion Degree, on PFS and EC. Pareto solution exploration for multi-objective optimization unveils the ideal urban configuration interval. To minimize PFS while maximizing EC, the research suggests pertinent measures: augmenting vegetation density, regulating the impervious coverage ratio within 50–70%, limiting two- and three-dimensional building density thresholds, and moderately escalating urban drainage network density. Additionally, it encourages a comprehensive appreciation of function-oriented land usage and intrinsic site topographical characteristics to reconcile varied urban development goals during planning. By fusing data-derived insights with multi-objective optimization, this research anticipates influencing urban planning models, thus enhancing decision-making related to urban configuration and fostering flood-resilient, sustainable, and economically prosperous urban habitats.

Multi kernel cross sparse graph attention convolutional neural network for brain magnetic resonance imaging super-resolution

High-resolution Magnetic Resonance Imaging (MRI) is pivotal in both diagnosing and treating brain tumors, assisting physicians in diagnosis and treatment by displaying anatomical structures. Utilizing convolutional neural network-based super-resolution methods enables the efficient acquisition of high-resolution MRI images.However, Convolutional neural networks are limited by their kernel size, which restricts their ability to capture a wider field of view, potentially leading to feature omission and difficulties in establishing global and local feature relationships. To overcome these shortcomings, We have designed a novel network architecture that highlights three main modules: (i) Multiple Convolutional Feature (MCF)extraction module, which diversifies convolution operations for extracting image features, achieving comprehensive feature representation. (ii) Multiple Groupss of Cross-Iterative Feature(MGCIF) modules, promoting inter-channel feature interactions and emphasizing crucial features needed for subsequent learning. (iii) A Graph Neural Network Module based on a sparse attention mechanism, capable of connecting distant pixel features and identifying influential neighboring pixels for target pixel inpainting. To evaluate the accuracy of our proposed network, we conducted tests on four datasets, comprising two sets of brain tumor data and two sets of healthy head MRI data, all of which underwent varying degrees of degradation. We conducted experiments using nineteen super-resolution (SR) models. Our experiments were carried out on four datasets, and the results demonstrate that our method outperforms the current leading-edge methods. In the four datasets, our model showed improvements in Peak Signal-to-Noise Ratio (PSNR) scores compared to the second-place model,with increase of 1.16 %,1.08 %,0.19 % and 0.53 % for ×2,and 2.26 %,1.67 %,0.13 % and 0.45 % for × 4,respectively.

Feasibility of low-cost particulate matter sensors for long-term environmental monitoring: Field evaluation and calibration

Low-cost sensor networks offer the potential to reduce monitoring costs while providing high-resolution spatiotemporal data on pollutant levels. However, these sensors come with limitations, and many aspects of their field performance remain underexplored. During October to December 2023, this study deployed two identical low-cost sensor systems near an urban standard monitoring station to record PM2.5 and PM10 concentrations, along with environmental temperature and humidity. Our evaluation of the monitoring performance of these sensors revealed a broad data distribution with a systematic overestimation; this overestimation was more pronounced in PM10 readings. The sensors showed good consistency (R2 > 0.9, NRMSE<5 %), and normalization residuals were tracked to assess stability, which, despite occasional environmental influences, remained generally stable. A lateral comparison of four calibration models (MLR, SVR, RF, XGBoost) demonstrated superior performance of RF and XGBoost over others, particularly with RF showing enhanced effectiveness on the test set. SHAP analysis identified sensor readings as the most critical variable, underscoring their pivotal role in predictive modeling. Relative humidity consistently proved more significant than dew point and temperature, with higher RH levels typically having a positive impact on model outputs. The study indicates that, with appropriate calibration, sensors can supplement the sparse networks of regulatory-grade instruments, enabling dense neighborhood-scale monitoring and a better understanding of temporal air quality trends.

GEOSPATIAL MONITORING OF THE ROAD NETWORK OF THE DNIPROPETROV REGION

The development of the road transport network is considered to be one of the crucial factors of the sustainable development of the country and regions. The presence of an extensive and topologically efficient road network and their appropriate technical condition ensure not only transport accessibility of settlements and road safety and stability, but also the speed and volume of transportation. The geographical location of Dnipropetrovsk oblast determines the important role of the region's road transport network in the formation of both external and intra-regional socio-economic relations. The oblast's road transport network consists of a few tens of thousands of kilometres of roads of various types of surfaces, and it has a complex internal structure, specific organization, and functional orientation. Therefore, geospatial monitoring of roads in Dnipropetrovsk oblast is an effective instrument for solving a number of regional transportation problems. On the basis of crowdsourced geospatial data from Open Street Map using QGIS, the study of spatial and territorial features of networks (road length, level of spatial complexity, structure, provision, stability, density, etc.) of different types of roads both for the oblast as a whole and over its administrative districts was conducted. The calculation of the total length of the region's roads using geospatial analysis showed a figure of 70823 km (all types of pavement), of which 3005 km are roads of national importance (4% of the total length), 33219 km are roads of local importance (47%) and 34599 km are roads in settlements (49%). It has been established that the structure of the state-owned road network in the oblast varies from a centralized type in the western (with a central hub in the city of Kryvyi Rih), northern (with a central hub in the village of Mahdalynivka) and north-eastern (with a central hub in the city of Pavlohrad) parts to a distributed type in other regions. It should be noted that the centralized structure of the western part of the region is formed by national and regional roads, the northern part by territorial roads, and the north-eastern part by international and territorial roads. This plays an important role in assessing the efficiency and sustainability of the network. The length of state roads of the territorial type in the region significantly dominates the length of international and national roads, which, given the European integration orientation of Ukraine's economy, indicates shortcomings in the state of the region's road transport network. At the same time, almost all international roads pass through the oblast's cities and towns, which does not meet the requirements for this type of road. Therefore, given the favourable geographical location of the oblast, the formation of a new network of highways through its territory will ensure intensive socio-economic development of both the region and Ukraine as a whole. The average density of the national highway network is only 0.0049 km/km². This is several times less than the rate used in developed countries for citywide highways. The hot spots on the map are connected with the intersections of roads of different levels and are mostly located outside the territories of settlements. This allows us to conclude that the road network framework is a compromise between the historical location of settlements, natural and geographical features of the territory (terrain, hydrology, geological structure, etc.) and construction cost constraints. The length of local roads is several times longer than the length of national roads. As a result, their average density for the region is 1.04 km/km², with the highest concentration in the central parts of the region and the lowest in the east and south of the region. The analysis of the region's roads by administrative units (raions) confirmed their geospatial diversity. The longest roads are observed in the most economically developed districts – Dnipro and Kryvyi Rih raions, and the shortest – in Pavlohrad and Nikopol raions. The same tendency is typical for local roads and roads in the settlements of the raions. However, the longest roads of national importance are in Synelnykove, Dnipro and Kamianske raions. All types of roads in the raions of the eastern and north-eastern parts of the oblast (Novomoskovsk, Pavlohrad, and Synelnykove) are characterized by the highest figures for total density, demonstrating the best transport provision of roads in Dnipropetrovsk oblast in contrast to the western districts. The main reason for this is the low population size in the east of the oblast. Keywords: geospatial monitoring, road network, network structure, transport availability, road density, Dnipropetrovsk region.

Lithological control of the active Mejillones fault from Triassic to Quaternary, northern Chile

Understanding the lithological control of crustal faults near cities is important for assessing the seismic hazard and characterising complex fault systems. We studied the Mejillones Fault (MF), located near one of the strategic ports in northern Chile. The MF is an active geological fault that extends 40 km in the N-S direction and is part of an extensive complex system of structures known as the Atacama Fault System (AFS). We built an S-wave tomography model using seismic ambient noise data that was recorded continuously in a dense network of stations from 2013 to 2015. The tomographic maps exhibited velocity anomalies that correlate with surface geology. We found lower velocities in the western part of the MF than in the eastern zone, with a shear wave velocity (Vs) of about 2.4 km/s, associated to fractured metamorphic rocks (Triassic) and marine sediments (Neogene-Quaternary). Higher velocities (Vs > 3.0 km/s) in the eastern part of the fault are related to igneous rocks from the Upper Jurassic. The difference in velocities between the western and eastern zones of the MF remained persistent in depth. The potential depth of the MF could be >5 km and delimits the boundary between 2 different geological zones. Intermediate Vs (∼2.8 km/s) at −23.25°, observed up to 4 km depth, can be associated with andesitic rocks deposited in an intra-arc basin (Upper Triassic) and accommodated by the MF over the last ∼200 Myr, evidencing the activity of this fault from Triassic to Quaternary.

Dynamic response and deformation failure microscopic characteristics of cyclic impact-cemented tailing backfill.

Frequent blasting disruptions can lead to cumulative damage within the cemented tailing backfill (CTB), increasing the risks associated with mining operations and reducing the recovery rate of the pillar. To address this issue, the Split Hopkinson Pressure Bar (SHPB) was utilized to conduct cyclic impact tests on CTB containing various cement tailing ratios (CTR) at different curing ages. The tests analyzed the stress-strain curve law, dynamic compressive strength (DCS), dynamic strength increase factor (DIF), absorption energy, and deformation failure characteristics of CTB under different impact velocities. Additionally, nuclear magnetic resonance (NMR) and scanning electron microscopy (SEM) were employed to investigate the internal pore structural properties of CTB. The research findings indicate that (1) Average strain rate exhibits a linear relationship with the DCS and impact velocity. A lower number of impacts occurred at higher impact velocities and shorter curing age. The number of impacts was drastically reduced when the impact velocity surpassed 3m/s. As the CTR increased, the number of impacts also increased. When the number of impacts increased, the elastic modulus, dynamic impact strength, and peak strain initially increased before ultimately decreasing. (2) Under the cyclic impact load, the shear failure and axial splitting failure were the main failure modes of CTB. Increasing the CTR may be a more effective strategy for reducing the degree of CTB fragmentation compared to prolonging the curing age. When the impact velocity is lower than 3m/s, CTB can withstand multiple impacts and maintain high levels of integrity. When the DIF of the first shock is below 1.5, the CTB demonstrates a capability to withstand more than four shocks. If the DIF exceeds 2, the CTB can only endure a single shock. (3) NMR and SEM observations revealed that CTB itself contains more pores. A dense network structure will grow inside CTB as the curing age and CTR are increased, reducing the porosity. The pore size observed in the samples also support that increasing CTR may be a more effective strategy. Our findings contribute to a better understanding of the kinetic response of CTB in deep mines under frequent blasting disruption and offer a valuable reference point for future research in this area.

Fe3+-modified polyamide membrane with a porous network structure: Rapid permeation and efficient interception of trace organic compounds

Forward osmosis (FO) is a promising technology in the treatment of trace organic substances. However, the development of this technology is limited by membrane material selection and the long-term chemical stability and durability of the membrane. Whereas ibuprofen, a representative trace organic substances that is hydrophobic and negatively charged, poses potential environmental hazards due to its bioactivity. In this work, polyethersulfone (PES) membranes were modified with iron metal complexes via chelation reactions to achieve high rejection rates for ibuprofen. The Fe3+-modified composite FO membrane increased the ibuprofen rejection rate to 98.3 % compared to less than 80.0 % for the original PES membrane. Additionally, the modified membrane enhanced water flux by at least 25.0 % and reduced reverse solute flux to a maximum of 73.0 % of that of the polyamide composite membrane, while significantly improving anti-fouling capabilities. Membrane characterization results revealed that the significant enhancement in the ibuprofen retention rate was due to the addition of a negative charge by forming complexes with Fe3+, which increased intermolecular interactions. Additionally, the dense network structure developed by the Fe3+-EDTA-2Na modification also contributed to the enhanced membrane performance.

Multivariate Hydrological Modeling Based on Long Short-Term Memory Networks for Water Level Forecasting

In the Department of Chocó, flooding poses a recurrent and significant challenge due to heavy rainfall and the dense network of rivers characterizing the region. However, the lack of adequate infrastructure to prevent and predict floods exacerbates this situation. The absence of early warning systems, the scarcity of meteorological and hydrological monitoring stations, and deficiencies in urban planning contribute to the vulnerability of communities to these phenomena. It is imperative to invest in flood prediction and prevention infrastructure, including advanced monitoring systems, the development of hydrological prediction models, and the construction of hydraulic infrastructure, to reduce risk and protect vulnerable communities in Chocó. Additionally, raising public awareness of the associated risks and encouraging the adoption of mitigation and preparedness measures throughout the population are essential. This study introduces a novel approach for the multivariate prediction of hydrological variables, specifically focusing on water level forecasts for two hydrological stations along the Atrato River in Colombia. The model, utilizing a specialized type of recurrent neural network (RNN) called the long short-term memory (LSTM) network, integrates data from hydrological variables, such as the flow, precipitation, and level. With a model architecture featuring four inputs and two outputs, where flow and precipitation serve as inputs and the level serves as the output for each station, the LSTM model is adept at capturing the complex dynamics and cross-correlations among these variables. Validation involves comparing the LSTM model’s performance with linear and nonlinear Autoregressive with Exogenous Input (NARX) models, considering factors such as the estimation error and computational time. Furthermore, this study explores different scenarios for water level prediction, aiming to utilize the proposed approach as an effective flood early warning system.

Characterizing collaborative networks for different arctic issues based on complex network analysis

International collaboration is widely accepted as an effective way to address Arctic-related issues, and these collaborative events have been frequently reported in recent years. To obtain the potential regular pattern hidden behind these events, a comprehensive methodology is proposed in this study. A database containing all the international collaborative events in terms of Arctic issues reported in the last decade was first developed, and these events were categorized into six aspects. Correspondingly, six subnetworks are developed, in which the nodes are represented by collaborative states, while the edges are denoted by collaborative relationships. These subnetworks are then characterized and compared by quantitative network parameters of total connections, network density, average path length, network transitivity, degree centrality, closeness centrality, and clustering coefficients. Finally, the performance of the individual states in each subnetwork is quantitatively analysed. The results show that Arctic science and research are the top priorities for international collaboration, and many uncertainties are involved in collaboration for Arctic resource utilization. In addition, the challenges faced by the non-Arctic states participating in Arctic issues are explained, and suggestions to promote international collaboration in terms of Arctic issues are proposed.

APTES/Calcium silica hydrate nano-hybrid composites with enhanced mechanical properties

Reducing the cement content in the concrete mix by developing cementitious materials with enhanced properties can contribute considerably to cut the CO2 emissions produced by the cement industry. In this work, we used a bottom-up approach to enhance the mechanical performance of the calcium silicate hydrate (C-S-H), by the covalent bonding of the 3-aminopropyltriethoxysilane (APTES), to create a nano-hybrid C-S-H/APTES. The intercalation of APTES within the C-S-H contributes to increase the bulk modulus by 35 % due to the stiffening along all the lattice directions. The stiffer behavior along a- and b- lattice directions is due to the formation of covalent bonds between the APTES molecules and the non-bonding oxygens of the C-S-H, blocking the silicate chain relaxation and increasing the overall stiffness of the ab-plane. In the c-direction, the stiffening is attributed to the densification of the hydrogen bond network in the interlaminar space caused by the APTES intercalation.

Network-based time series modeling for COVID-19 incidence in the Republic of Ireland

Network-based time series models have experienced a surge in popularity over the past years due to their ability to model temporal and spatial dependencies, arising from the spread of infectious disease. The generalised network autoregressive (GNAR) model conceptualises time series on the vertices of a network; it has an autoregressive component for temporal dependence and a spatial autoregressive component for dependence between neighbouring vertices in the network. Consequently, the choice of underlying network is essential. This paper assesses the performance of GNAR models on different networks in predicting COVID-19 cases for the 26 counties in the Republic of Ireland, over two distinct pandemic phases (restricted and unrestricted), characterised by inter-county movement restrictions. Ten static networks are constructed, in which vertices represent counties, and edges are built upon neighbourhood relations, such as railway lines. We find that a GNAR model based on the fairly sparse Economic hub network explains the data best for the restricted pandemic phase while the fairly dense 21-nearest neighbour network performs best for the unrestricted phase. Across phases, GNAR models have higher predictive accuracy than standard ARIMA models which ignore the network structure. For county-specific predictions, in pandemic phases with more lenient or no COVID-19 regulation, the network effect is not quite as pronounced. The results indicate some robustness to the precise network architecture as long as the densities of the networks are similar. An analysis of the residuals justifies the model assumptions for the restricted phase but raises questions regarding their validity for the unrestricted phase. While generally performing better than ARIMA models which ignore network effects, there is scope for further development of the GNAR model to better model complex infectious diseases, including COVID-19.

High performance magnesium-based plastic semiconductors for flexible thermoelectrics

Low-cost thermoelectric materials with simultaneous high performance and superior plasticity at room temperature are urgently demanded due to the lack of ever-lasting power supply for flexible electronics. However, the inherent brittleness in conventional thermoelectric semiconductors and the inferior thermoelectric performance in plastic organics/inorganics severely limit such applications. Here, we report low-cost inorganic polycrystalline Mg3Sb0.5Bi1.498Te0.002, which demonstrates a remarkable combination of large strain (~ 43%) and high figure of merit zT (~ 0.72) at room temperature, surpassing both brittle Bi2(Te,Se)3 (strain ≤ 5%) and plastic Ag2(Te,Se,S) and organics (zT ≤ 0.4). By revealing the inherent high plasticity in Mg3Sb2 and Mg3Bi2, capable of sustaining over 30% compressive strain in polycrystalline form, and the remarkable deformability of single-crystalline Mg3Bi2 under bending, cutting, and twisting, we optimize the Bi contents in Mg3Sb2-xBix (x = 0 to 1) to simultaneously boost its room-temperature thermoelectric performance and plasticity. The exceptional plasticity of Mg3Sb2-xBix is further revealed to be brought by the presence of a dense dislocation network and the persistent Mg-Sb/Bi bonds during slipping. Leveraging its high plasticity and strength, polycrystalline Mg3Sb2-xBix can be easily processed into micro-scale dimensions. As a result, we successfully fabricate both in-plane and out-of-plane flexible Mg3Sb2-xBix thermoelectric modules, demonstrating promising power density. The inherent remarkable plasticity and high thermoelectric performance of Mg3Sb2-xBix hold the potential for significant advancements in flexible electronics and also inspire further exploration of plastic inorganic semiconductors.

Russian humanitarian policy: problems and solutions

The article examines some problems of Russia’s humanitarian policy and humanitarian cooperation. According to the author, domestic humanitarianism does not go beyond the boundaries established by the Helsinki Final Act. The formulation of tasks and directions of domestic humanitarian policy contained in strategic planning documents is reduced only to a description of the methods and conditions of communication, leaving the specifics of humanitarian activities and humanitarian instruments outside the scope of strategic planning documents, thereby complicating its institutionalization. The author believes that the goal of humanitarian cooperation may be the formation of a global structure of support for domestic foreign policy initiatives, representing a dense network of interactions that provides ripple effects, expanding Russia’s influence and forming new dependencies. But achieving such a goal is impossible without putting forward a clear alternative to the modern world order, which will reflect not only the interests of Russia, but also of countries that disagree with the Western-centric world order.

Weight-adaptive joint mixed-precision quantization and pruning for neural network-based equalization in short-reach direct detection links.

Neural network (NN)-based equalizers have been widely applied for dealing with nonlinear impairments in intensity-modulated direct detection (IM/DD) systems due to their excellent performance. However, the computational complexity (CC) is a major concern that limits the real-time application of NN-based receivers. In this Letter, we propose, to our knowledge, a novel weight-adaptive joint mixed-precision quantization and pruning approach to reduce the CC of NN-based equalizers, where only integer arithmetic is taken into account instead of floating-point operations. The NN connections are either directly cutoff or represented by a proper number of quantization bits by weight partitioning, leading to a hybrid compressed sparse network that computes much faster and consumes less hardware resources. The proposed approach is verified in a 50-Gb/s 25-km pulse amplitude modulation (PAM)-4 IM/DD link using a directly modulated laser (DML) in the C-band. Compared with the traditional fully connected NN-based equalizer operated with standard floating-point arithmetic, about 80% memory can be saved at a minimum network size without degrading the system performance. Quantization is also shown to be more suitable to over-parameterized NN-based equalizers compared with NNs selected at a minimum size.

A Hybrid Anti-Collision Protocol Based on Frequency Division Multiple Access (FDMA) and Time Division Multiple Access (TDMA) for Radio Frequency Identification (RFID) Readers

Radio Frequency Identification (RFID) technology plays a crucial role in various Internet of Things (IoT) applications, necessitating the integration of RFID systems into dense networks. However, the presence of numerous readers leads to collisions, degrading communication between readers and tags and compromising system performance. To tackle this challenge, researchers have proposed Medium Access Control (MAC) layer protocols employing different channel access methods. In this paper, we present a novel solution, the Distributed Time Slot Anti-Collision protocol (DTS-AC), which employs a new TDMA notification system to address Reader-to-Reader Interference (RRI), while incorporating FDMA-based frequency resource management to resolve Reader-to-Tag Interference (RTI) collision issues. Simulation results demonstrate that DTS-AC significantly improves performance in dense RFID networks by enhancing read rates, with scalability benefits based on the number of readers, channels, and Time Slots (TSs). Moreover, the cost-effectiveness of DTS-AC facilitates efficient deployment in RFID networks, emphasizing considerations of time delay and data sensitivity.

Improvement of O/W emulsion stability and rheological properties by highly branched cyclic dextrin and preparation of a novel emulsion gel

The objective of this study was to formulate and investigate a low-oil emulsion gel (20%, v/v) stabilized by highly branched cyclic dextrin (HBCD). The results indicated that HBCD effectively prevented the coalescence of oil droplets when the concentration reached a specific threshold (40 wt%). All samples exhibited dense network structure and elastic solid properties when stored at 4 °C. Fluorescence images showed the oil droplets tended to be irregular after refrigeration, indicating HBCD molecules aggregated under low temperature. Fourier transform infrared spectroscopy and its deconvolution results displayed a shift in the hydroxyl peak and an increase in R1046/1022. Additionally, new crystalline peaks emerged at 17° and 22° in the X-ray diffusion spectra of recrystallized HBCD samples. It suggested the hydrogen bonds and ordered structure of HBCD were increased in the emulsion gels during storage. Therefore, the formation and orderly arrangement of double helix structures of the branched chains which stabilized by hydrogen bonds promoted the recrystallization and self-assembly of HBCD, ultimately resulting in the formation of a stable gel structure. The research provided theoretical feasibility for the development of structured emulsions for functional foods.

How does the built environment affect the usage efficiency of dockless-shared bicycle? An exploration of time-varying nonlinear relationships

The imbalance in Dockless Bike Sharing (DBS) systems is a major concern for planners, causing a significant drop in utilization efficiency. However, limited research quantifies DBS usage efficiency from a supply-demand perspective, also, the understanding of the nonlinear relationship between the built environment and DBS utilization efficiency from the time dimension is lacking, leading to biased assessments and the losses of flexible and effective DBS rebalancing strategies. Therefore, this study quantifies the efficiency of DBS usage from a supply-demand perspective by calculating the average usage interval of DBS facilities within urban subzones, termed duration of stopping usage (DSU), and employs emerging eXplainable Artificial Intelligence (XAI) technology to reveal the time-varying nonlinear impact of the built environment on DSU. The results show that the relative importance of transit accessibility, land use mix entropy and road network density remains stable in the time dimension. The time non-stationarity of the nonlinear relationship between these variables and DSU is primarily manifested in dynamic shifts of thresholds. Notably, the time-varying nature of the relative importance is particularly prominent for variables related to land use facilities. Moreover, the time non-stationarity of the nonlinear relationship is more complex, manifesting not only in threshold shifts but also in changes in correlation. We also propose several spatial transfer methods for DBS facilities, offering fresh insights for crafting flexible and adaptive DBS rebalancing strategies. These findings enhance the interpretability of the inconsistent impact of the built environment on DBS utilization efficiency and provide valuable knowledge for scientific management decisions regarding DBS rebalancing.

Salt thickening performance and mechanism of an N-Vinyl-2-Pyrrolidinone based amphiphilic polymer

Enhancing salt tolerance is a critical requirement for polymer flooding in the petroleum production process. This work introduced a novel salt thickening polymer called NDC16, which is a cationic polymer composed of poly(N-Vinyl-2-Pyrrolidinone-co-Double hexadecyldimethylallyl ammonium chloride). Notably, NDC16 does not contain traditional Acrylamide (AM) or Acrylic acid (AA). Its salt thickening performance and mechanism in sodium chloride and calcium chloride solutions were separately analyzed at aggregation and molecular levels by photophysical method, micromorphology, and molecular dynamics (MD) simulation. The results demonstrated that the novel polymer was able to increase the viscosity of the sodium chloride solution to 20 times higher than the initial one. However, its performance in increasing viscosity in a calcium chloride solution was only half as effective. The inclusion experiments indicated that almost all increasing viscosity could be contributed to the hydrophobic association effects of the polymer molecules. The results of the dynamic analysis revealed that the NDC16 aggregates showed a more orderly arrangement in NaCl solutions, while the network structure appeared as relatively disordered clusters in CaCl2. Besides, there existed an optimal salinity at which the network density of the hydrophobic structure reached the highest level. The variation in thickening performance can be elucidated by the molecular interaction. Due to their smaller molecular interactions, the Na+ ions exhibited a stronger potential to permeate into the polymer clusters. It stretched the polymer molecules and allowed them to display a structured arrangement. The large Radius of Gyration (Rg) and spatial distance of the polymers in two solutions indicated that the polymer clusters were much elongated in Na+ solvents, thereby confirming their permeation effect. These results enhance the comprehension of the process by which salt thickening occurs due to the interaction between cationic hydrophobic associative polymers and various salt solutions.

Comparison of Deep and Machine Learning Approaches for Quebec Tree Species Classification Using a Combination of Multispectral and LiDAR Data

Accessing tree species information is required for making appropriate decisions in forest management. 3D photo-interpretation using high spatial resolution aerial imagery is used to provide information on tree species in the province of Quebec. However, the shortage of qualified interpreters and the increasing costs of 3D photo-interpretation have affected the production of the forest inventory. In this study, we employed deep and machine learning models to classify nine tree species (i.e., paper birch, yellow birch, red maple, poplar, black spruce, white spruce, tamarack, jack pine, and balsam fir). We used a combination of spectral and vertical structural information extracted from 30-cm aerial imagery and airborne light detection and ranging (LiDAR) data in a 9 , 100 km 2 forested area in Quebec, Canada. The results indicated that Dense Convolution Network (DenseNet) achieved the best overall accuracy of 78%, outperforming machine learning methods by 5%. In addition, the models’ performance was independently assessed using permanent and temporary ground sample plots, acknowledging the superiority of DenseNet in terms of overall accuracy (73%) in predicting the dominant species. Our results suggest that the combination of aerial imagery and airborne LiDAR data, using deep learning approaches, can be applied to accurately map tree species.

Efficient construction and convergence analysis of sparse convolutional neural networks

In this paper, a new variant of convolutional neural network (CNN) based on L1 regularization is proposed. The main consideration is to generate a sparse weight matrix, i.e., to generate a sparse neural network model. To some extent, L1 regularization prevents the occurrence of overfitting and effectively improves the learning efficiency of the network. Unfortunately, the L1 regularization term is non-smooth, which leads to the occurrence of oscillations in experiments and poses a great challenge to the theoretical analysis. So, we overcome this drawback by using polishing techniques and rigorously prove the monotonicity of the error function and the strong and weak convergence theorems of the algorithm. Finally, numerical simulations on several data sets support our theoretical results and the superiority of the proposed algorithm.