Transfer Analysis Research Articles

Chaotic neural network quantization and its robustness against adversarial attacks

Achieving robustness against adversarial attacks while maintaining high accuracy remains a critical challenge in neural networks. Parameter quantization is one of the main approaches used to compress deep neural networks to have less inference time and less storage memory size. However, quantization causes severe degradation in accuracy and consequently in model robustness. This work investigates the efficacy of stochastic quantization to enhance robustness and accuracy. Noise injection during quantization is explored to understand the impact of noise types and magnitudes on model performance. A comprehensive comparison between different applying scenarios for stochastic quantization and different noise types and magnitudes was implemented in this paper. Compared to the baseline deterministic quantization, chaotic quantization achieves a comparable accuracy, however, it achieves up to a 43% increase in accuracy against various attack scenarios. This highlights stochastic quantization as a promising defense mechanism. In addition, there is a crucial role played by the choice of noise type and magnitude in stochastic quantization. Lorenz and Henon noise distributions in stochastic quantization outperform traditional uniform and Gaussian noise in defending against attacks. A transferability analysis was discussed to understand the generalizability and effectiveness of the proposed stochastic quantization techniques. A cross-validation definition was newly evaluated in this scope to analyse the model’s stability and robustness against attacks. The study outperformed a quantization network technique and improved the model’s robustness and stability against adversarial attacks using chaotic quantization instead of deterministic quantization or even instead of stochastic quantization using traditional noise.

Hydrocephalus in the emergency department: descriptive nationwide data and analysis of transfer.

Hydrocephalus is a chronic, treatable, but in most cases incurable condition characterized by long periods of stability punctuated by crises. Patients in crisis usually seek care in an emergency department (ED). How patients with hydrocephalus use EDs has received almost no epidemiological study. Data were taken from the National Emergency Department Survey for 2018. Visits by patients with hydrocephalus were identified by diagnostic codes. Neurosurgical visits were identified by codes for imaging of the brain or skull or by neurosurgical procedure codes. Visits and dispositions were characterized by demographic factors for neurosurgical and unspecified visits by using methods for analysis of complex survey designs. Associations among demographic factors were assessed using latent class analysis. There were an estimated 204,785 ED visits by patients with hydrocephalus in the United States in 2018. Roughly 80% of patients with hydrocephalus who visited EDs were adults or elders. By a ratio of 2:1, patients with hydrocephalus visited EDs much more often for unspecified reasons than for neurosurgical reasons. Patients with neurosurgical complaints had more costly ED visits, and if they were admitted they had longer and more costly hospitalizations than did patients with unspecified complaints. Only 1 in 3 patients with hydrocephalus who visited an ED was sent home regardless of whether the complaint was neurosurgical. Neurosurgical visits ended in transfer to another acute care facility more than 3 times as often as unspecified visits. Odds of transfer were more strongly associated with geography and, specifically, with proximity to a teaching hospital than with personal or community wealth. Patients with hydrocephalus make heavy use of EDs, and they make more visits for reasons unrelated to their hydrocephalus than for neurosurgical reasons. Transfer to another acute care facility is an adverse clinical outcome that is much more common after neurosurgical visits. It is a system inefficiency that might be minimized by proactive case management and coordination of care.

AS2T: Arbitrary Source-To-Target Adversarial Attack on Speaker Recognition Systems

Recent work has illuminated the vulnerability of speaker recognition systems (SRSs) against adversarial attacks, raising significant security concerns in deploying SRSs. However, they considered only a few settings (e.g., some combinations of source and target speakers), leaving many interesting and important settings in real-world attack scenarios alone. In this work, we present <small>AS2T</small> , the first attack in this domain which covers all the settings, thus allows the adversary to craft adversarial voices using arbitrary source and target speakers for any of three main recognition tasks. Since none of the existing loss functions can be applied to all the settings, we explore many candidate loss functions for each setting including the existing and newly designed ones. We thoroughly evaluate their efficacy and find that some existing loss functions are suboptimal. Then, to improve the robustness of <small>AS2T</small> towards practical over-the-air attack, we study the possible distortions occurred in over-the-air transmission, utilize different transformation functions with different parameters to model those distortions, and incorporate them into the generation of adversarial voices. Our simulated over-the-air evaluation validates the effectiveness of our solution in producing robust adversarial voices which remain effective under various hardware devices and various acoustic environments with different reverberation, ambient noises, and noise levels. Finally, we leverage <small>AS2T</small> to perform thus far the largest-scale evaluation to understand transferability among 14 diverse SRSs. The transferability analysis provides many interesting and useful insights which challenge several findings and conclusion drawn in previous works in the image domain. Our study also sheds light on future directions of adversarial attacks in the speaker recognition domain.

Spatial Transferability of Residential Building Damage Models between Coastal and Fluvial Flood Hazard Contexts

This study investigates residential building damage model transferability between coastal and fluvial flood hazard contexts. Despite the frequency of damaging coastal flood events, empirical damage models from fluvial flooding are often applied in quantitative coastal flood risk assessments. This assumes that building damage response is similar from the exposure to different flood sources. Here, we use empirical data from coastal, riverine and riverine-levee breach flooding events to analyse residential building damage. Damage is analysed by applying univariable and multivariable learning models to determine the importance of explanatory variables for relative damage prediction. We observed that the larger explanatory variable range considered in multivariable models led to higher predictive accuracy than univariable models in all flood contexts. Transfer analysis using multivariable models showed that models trained on event-specific damage data had higher predictive accuracy than models learned on all damage data or on data from other events and locations. This finding highlights the need for damage models to replicate local damage factors for reliable application across different flood hazard contexts.

Landscape ecological risk assessment of chongming dongtan wetland in shanghai from 1990 to 2020

Landscape ecological risk assessment can assess the stability, vulnerability, and anti-interference ability of the ecosystem based on the characteristics, structure, and function of the landscape in order to determine the degree and distribution of landscape ecological risk. In recent years, due to human activities and global climate change, the area and ecological quality of the Chongming Dongtan wetland have been decreasing. As an important wintering habitat and transit point on the ‘East Asia-Australia migration line’, the population and habitat security of wintering migratory birds are also threatened. Therefore, the landscape index method was used to assess the ecological risk of the Chongming Dongtan wetland. Based on transfer analysis and spatial autocorrelation analysis, the spatial and temporal changes of landscape ecological risk were revealed. The findings are as follows: (1) Natural landscapes such as mudflats, grasslands, and waters have been declining, while artificial landscapes such as cultivated land, construction land, and woodland have been gradually expanding, indicating a shift from natural to artificial landscapes. (2) Overall, the landscape ecological risk is showing an increasing trend. Wetlands have moved from being dominated by the lowest and medium ecological risks to being evenly distributed by all ecological risk classes. Notably, there has been a rapid increase in landscape ecological risk in 2000. The highest-risk areas have expanded by 9346.95 hm2, gradually encroaching from the buffer zone into the reclamation area and the nature reserve. Conversely, the lowest ecological risk areas have decreased by 7091.82 hm2, primarily concentrated in nature reserves. (3) The aggregation of landscape ecological risks continues to decrease, indicating a reduced likelihood of dramatic changes in ecological risks.

Object relations theory and transference analysis

Object relations theory and transference analysis

Transference-Focused Psychotherapy for Adolescents with Personality Disorders: A Case Example with a Focus on the Interpretative Process and Transference Analysis

ABSTRACT In this paper, I focus on the nature and functions of the interpretive process in psychotherapy with adolescents with severe personality disorder. I will describe how the interpretive process in transference focused psychotherapy for adolescents provides affective containment and symbolic management of psychological conflict along with a shift toward the predominance of repression-based over primitive defenses. This process leads to a progressive resolution of identity diffusion. I will give illustrations of different phases of the interpretive process.

Inside the Body: How the Science of Interoception Can Provide a New Understanding of Psychoanalysis

ABSTRACT Our brains’ ability to attend to our internal body sensations (“interoception”) safeguards our physiological equilibrium and, in so doing, creates our emotional awareness, sense of well-being, and self-experience. To my mind, recent interoception research syncs nicely with basic psychoanalytic theory. It offers a potentially revelatory new understanding of how the analysis of transference “works” to create change. The “predictive coding” model probes the nitty-gritty of how our brains create negative, fearful expectations based on past experience and how these negative predictions actually create the feared negative outcome. Powerful early emotional experiences, which cannot be consciously remembered per se, continue to exist mainly as felt bodily sensations. The only way to access such early patterns (“predictions”) is through accessing somatic sensation. We must, therefore, explicitly help our patients learn to identify, reflect upon and verbalize internal bodily signals, so that early dissociated emotional experience, “stored” in the body, can be brought to the level of subjective experience. The more interoceptively aware we are, the more we can filter our patients’ communications through our own interoceptive awareness, and the more effectively we can help our patients “update” their predictions and enliven their emotional lives.

Interhospital Patient Transfers from the Emergency Department: A Retrospective, Observational Study

Aims: All patients requiring emergency intervention are given their first medical attention in the emergency department, regardless of the facilities, equipment, and occupancy of the hospital. When the hospital facilities are not sufficient, interhospital patient transfers may be done under some regulations and legislations. This study aimed to analyze interhospital patient transfers from emergency departments to other hospitals. Materials and methods: This retrospective, observational, descriptive study examined interhospital patient transfers from İstanbul Emergency Department to other hospitals for four years (2018–2022). While all referrals made by the emergency service through the Arnavutköy State Hospita 112 Provincial Ambulance Service Transport Unit were eligible for the study, patients under the age of 18, those whose referral was due to COVID-19 infection, and those whose data could not be accessed were excluded. In addition to demographic data, interhospital patient transfers were evaluated according to time, department, diagnosis, distance between hospitals, and the characteristics of the referral institution. Results: The study included 4,280 transfers. Transfers were done most frequently in 2018 (1,310; 30.6%). Transfers were often handled outside of weekday daytime (2,919; 68.2%) and were most frequently done for cardiac events (2,592; 60.6%). The average transfer distance between hospitals was calculated as 24.29 km (standard deviation [SD]: 6.55 km), while the mean time between the first registration and the transfer was 344.8 minutes (SD: 275.3 minutes). In the analysis of transfers by year, a higher rate of transfers were to public institutions in each successive year (2018: 40.8%; 2019: 52.4%; 2020: 77.8%; 2021: 78.3%). Conclusion: The most common reason for transfer was cardiac events. Increasing the cardiology facilities and equipment of the hospital could reduce the number of interhospital patient transfers.

Role of copper and alumina for heat transfer in hybrid nanofluid by using Fourier sine transform

The convection, thermal conductivity, and heat transfer of hybrid nanofluid through nanoparticles has become integral part of several natural and industrial processes. In this manuscript, a new fractionalized model based on hybrid nanofluid is proposed and investigated by employing singular verses and non-singular kernels. The mathematical modeling of hybrid nanofluid is handled via modern fractional definitions of differentiations. The combined Laplace and Fourier Sine transforms have been configurated on the governing equations of hybrid nanofluid. The analytical expression of the governing temperature and velocity equations of hybrid nanofluid have been solved via special functions. For the sake of thermal performance, dimensional analysis of governing equations and suitable boundary conditions based on Mittage-Leffler function have been invoked for the first time in literature. The comparative analysis of heat transfer from hybrid nanofluid has been observed through Caputo-Fabrizio and Atangana-Baleanu differential operators. Finally, our results suggest that volume fraction has the decelerated and accelerated trends of temperature distribution and inclined and declined profile of heat transfer is observed copper and alumina nanoparticles.

Three-dimensional analysis of heat transfer and entropy production of jet impingement hybrid nanofluid cooling a porous media-filled heat sink

ABSTRACT In this study, a numerical assessment was carried out on heat transfer and entropy production of a three-dimensional heat sink exposed to an impinging jet, where the area between the heat sink fins is filled with aluminum foam saturated with A L 2 O 3 − C u / w a t e r hybrid nanofluid. Darcy-Brinkman-Forchheimer’s model with the local thermal equilibrium was considered. The effects of various important parameters such as Reynolds number 400 ≤ R e ≤ 1600 , volumetric concentrations 1 % ≤ φ h n f ≤ 5 % , Darcy number 10 − 5 ≤ D a ≤ 10 − 1 , porosity 0.1 ≤ ε ≤ 0.6 , and the diameter of nanoparticles 20 n m ≤ d n ≤ 60 n m on fluid flow, heat transfer, drop pressure, and entropy generation were investigated. A finite-volume approach with SIMPLE algorithm was employed to solve continuity, momentum, energy, and entropy equations with the help of the Ansys-Fluent 14.5 program. Regarding the validation of the computational approach used in this paper, a strong consensus has been reached with other findings available in the literature. Outcomes indicated that boosting Darcy number D a and porosity ε of the used porous material helps increase the heat exchange rate by up to 72 % and 40 % , respectively. On the other hand, we obtained a 7 % enhancement in the heat transmission rate when using nanoparticles with a diameter d n of 20 n m instead of nanoparticles with a diameter of 60 nm. Furthermore, it has been demonstrated that boosting D a , ε and d n minimize drop pressure by various proportions. Based on the entropy production analysis, it is possible to infer that entropy due to heat transfer is more dominant than other types of entropy with all parameters analyzed. Also, it can be deducted that as the values of ε and D a grow and the values of d n decline, the total entropy production ( S p , t ) rises up to 55 % , 75 % and 3 % , respectively. The same findings were reached with the Bejan number B e , but different percentages. Finally, three correlations of Nusselt number with several variables with an error not exceeding 15 % have been suggested. The proposed study’s findings may be utilized to define appropriate process factors for improved thermodynamic efficiency of heat sinks using hybrid nanofluids and porous media.

Heat Transfer and Exergy Analysis of a Shell and Tube Heat Exchanger using PGW based ZnO Nanofluids

In this experimental work, ZnO nanoparticles were synthesized using the chemical precipitation method, and the nanoparticle structure and morphology were characterized through XRD and SEM. Heat transfer and exergetic characteristics were then studied in a shell and tube heat exchanger using PGW-based ZnO nanofluids varying nanoparticle volume concentration and nanofluid (shell side) flow rate at 6, 8, 10 and 12 litres/min. The hot water flow rate was fixed at 12 litres/min. The experimental results show that the heat transfer rate was improved by increasing the nanoparticle concentration and nanofluid flow rate. When the nanoparticle volume concentration was 0.3 per cent, the maximum enhancement of heat transfer rate and average heat transfer coefficient using ZnO nanofluids were 35.9 per cent and 40.2 per cent, respectively, in comparison to the base fluid. Exergy loss and dimensionless exergy loss both increased with nanofluid flow rate and dropped substantially with increased nanoparticle volume concentrations. The average increment of exergetic effectiveness at three different nanoparticle volume concentration (0.1%, 0.2%, and 0.3%) are 10.68%, 23.64%, and 31.23% respectively. The highest exergetic sustainability index (0.41) and lowest environmental impact factor (2.42) were observed when the nanoparticle concentration was 0.3% with the nanofluid flow rate of 6 litres/min.

Analysis of Industrial Carbon Transfer in Beijing-Tianjin-Hebei City Cluster and Surrounding Areas

To achieve the goal of carbon neutrality and win the blue-sky defense battle, the environmental situation in Beijing-Tianjin-Hebei and surrounding areas is still grim, and the optimization of its industrial structure and energy structure is imminent. With the rapid development of interregional trade in intermediate products, carbon emissions are transferred across regions with the trade. Due to the large differences in the technology, industrial structure, and economic development of cities, extending the environmental governance chain of Beijing-Tianjin-Hebei and surrounding areas is indispensable. In this article, based on the interregional input-output tables in 2002, 2007, and 2012, we establish the average propagation length (APL) model and the structural path analysis model Structural Path Analysis model for analyzing the carbon conduction relationship in Beijing-Tianjin-Hebei. And we also compare the situation of the Yangtze River Delta and the Pearl River Delta. The results show that: i) From perspective of the whole urban clusters, Beijing-Tianjin-Hebei has obvious characteristics of coal-fired urban clusters. More than 65% of the carbon-containing resources in Hebei’s coal industry are transferred to the electricity and heat industry. In the carbon conduction chain, the carbon emissions caused by electricity and heat industry, which acts as an intermediary, account for more than 85% of the total emissions. ii) From the perspective of industrial structure transfer within the urban clusters, Hebei Province has an important resource support position. Its secondary industry can not only effectively alleviate the shortage of energy supply in other resource provinces, but also has great development potential in the improvement of economic benefits. iii) From the perspective of specific industry sectors, resource provinces such as Shanxi and Inner Mongolia have high carbon emission coefficients in the electricity and heat industry, which is the main reason for the high carbon emissions in the transfer chain.

Influence Analysis of Convective Heat Transfer of Circular Tubes and Flat Plates in a Polar Environment

Electric heating serves as a major cold-proof measure because engineering equipment often loses efficacy in harsh weather conditions in a polar environment. Because Convective heat transfer is the key issue to solve the cold-proof design, this paper analyzes convective heat transfer of tube and plate in the constant heat flow heating mode. According to fluent numerical simulation analysis with a wind speed of 0–40 m/s and a temperature of −40°C–0°C, both an increase in wind speed and a decrease in temperature would cause the growth of the convective heat transfer coefficient of circular tubes and flat plates, Wind speed displayed a more significant impact on convective heat transfer. In a low-temperature laboratory, a model experiment of convective heat transfer was carried out to validate the reliability of the numerical simulation. Last, a prediction model for convective heat transfer coefficient of circular tubes and flat plates in polar environment was developed using numerical and experimental data.

Pd-GaSe and Pd3-GaSe Monolayers: Two Promising Candidates for Detecting Dissolved Gases in Transformer Oil

In this paper, the adsorption behaviors of three gases (H2, CO, and C2H2) decomposed by the transformer oil on Pd-GaSe and Pd3-GaSe monolayers were calculated by density functional theory. Compared with Pd single-atom doping, Pd3 cluster doping changed the original structure and charge distribution to a greater extent, and more obviously improved the conductivity. According to the analysis of adsorption energy, charge transfer and deformation charge density, the results show that the two doped structures have better adsorption performance for the three gas molecules (H2, CO, and C2H2) than the intrinsic GaSe monolayer. Compared with Pd-GaSe, Pd3-GaSe showed stronger adsorption property for the three gases. Analysis of frontier molecular orbitals and recovery characteristics shows that Pd3-GaSe can be used as an ideal gas sensitive material for H2 detection because of its good desorption properties and obvious conductivity changes. Pd-GaSe can be used as a disposable resistive sensor for CO. Pd3-GaSe is a kind of sensing material suitable for disposable resistance sensors for CO and C2H2. These two doped structures have great application potential in gas adsorption and detection, and provide indications for further study on gas sensor detection by means of metal-doped GaSe monolayer.

Three-dimensional numerical analysis of flow and heat transfer of bi-directional stretched nanofluid film exposed to an exponential heat generation using modified Buongiorno model

The heat transfer characteristics of copper/water nanofluid flow over a bi-directional stretched film are theoretically studied. The used mathematical model accounts for nanofluid effective dynamic viscosity and thermal conductivity. The model of the current study utilizes the modified Buongiorno model to scrutinize the effect of haphazard motion, nanoparticles' thermo-migration, and effective nanofluid properties. 3D flow is driven by having the nanofluid film elongation in two directions. The thermal analysis of the problem considers the nonlinear internal heat source and Newton heating conditions. In modeling the problem, the Prandtl boundary layer approximations are employed. Moreover, the nonlinear problem set of governing equations for investigating the transport of water conveying copper nanoparticles was non-dimensionalized before being treated numerically. The current parametric study investigates the impact of governing parameters on nanoparticles velocities, temperature, and concentration distributions. The presence of copper nanoparticles leads to a higher nanofluid temperature upon heating. The temperature enhances with the nanoparticles Brownian movement and thermo-migration aspects. Furthermore, involving a heat source phenomenon augments the magnitude of the heat transfer rate. Moreover, the velocity ratio factor exhibits decreasing behavior for x-component velocity and increasing behavior for y-component velocity. In conclusion, the study results proved that for larger values of Nb and Nt the temperature is higher. In addition, it is clear from the investigations that the Lewis number and Brownian motion factor decline the nanoparticle concentration field.

Dependence of Temperature Rise on the Position of Catheters and Implants Power Sources Due to the Heat Transfer into the Blood Flow

This work provides a numerical analysis of heat transfer from medical devices such as catheters and implants to the blood flow by considering the relative position of such power sources to the vessel wall. We have used COMSOL Multiphysics® software to simulate the heat transfer in the blood flow, using the finite element method and Carreau-–Yasuda fluid model (a non-Newtonian model for blood flow). The location of the power source is changed (from the center to near the wall) in the blood vessel with small steps, while the blood flow takes different velocities. The numerical simulations show that when the catheter/implant approaches the vessel wall, the temperature increases linearly for ~90% of the radial displacement from the centerline position to the vessel wall, while for the last 10% of the radial displacement, the temperature increases exponentially. As a result, the temperature is increased significantly, when changing the position of the catheter/implant from the centerline to the area adjacent to the vessel wall.

Coupled Analysis of Heat Transfer in a Molten Pool With Three-Layer Configuration

Understanding a reliable description of the heat transfer in the molten pool with three-layer configuration is inevitable for the design of severe accident mitigation measures in research reactors, such as IVR-ERVC. However, investigations on a thick lower metallic layer are scarce. Thus, further studies on heat transfer in a molten pool with a thick lower metallic layer are highly required, based on previous severe accident studies in light water reactors. This study conducts a numerical simulation of heat transfer in the molten pool with a thick lower metallic layer during a severe accident, combining model development, code verification, and code application. First, the MPCAP code has been developed, based on the coupled simulation of natural convection inside the molten pool, two-dimensional heat conduction in the RPV lower head, and convection outside the reactor vessel wall. Then, a verification case has been conducted using three-layer configuration in AP1000. The AP1000 verification case indicates that the MPCAP code predicts important parameters in the molten pool well.

Lie symmetry analysis of heat transfer in a liquid film over an unsteady stretching surface with viscous dissipation and external magnetic field

Fluid flow and heat transfer in thin films over unsteady stretching sheets are studied and analyzed by employing similarity transformations and analytical or numerical solution schemes. We derive Lie point symmetries for system of partial differential equations describing the flow and heat transfer in a thin liquid film on an unsteady stretching surface with viscous dissipation in the presence of external magnetic field. We deduce a 5-dimensional Lie point symmetry algebra. We use a couple of the admitted Lie point symmetry generators to construct new similarity transformations for the considered model, through associated invariants. These transformations enable reductions of the model to nonlinear ordinary differential equations. We construct analytic solutions for this system using the Homotopy analysis method to study the considered MHD flow and heat transfer. These results are presented in the form of tables and figures to reflect variations in fluid velocity, film thickness and heat transfer with the magnetic parameter, Prandtl number, Eckert number and unsteadiness parameter. We show that these variations in the velocity and temperature profiles are different from those reported earlier for the flow and heat transfer within a thin film under the effects of viscous dissipation and external magnetic field.

Towards autonomous analysis of chemical exchange saturation transfer experiments using deep neural networks

Macromolecules often exchange between functional states on timescales that can be accessed with NMR spectroscopy and many NMR tools have been developed to characterise the kinetics and thermodynamics of the exchange processes, as well as the structure of the conformers that are involved. However, analysis of the NMR data that report on exchanging macromolecules often hinges on complex least-squares fitting procedures as well as human experience and intuition, which, in some cases, limits the widespread use of the methods. The applications of deep neural networks (DNNs) and artificial intelligence have increased significantly in the sciences, and recently, specifically, within the field of biomolecular NMR, where DNNs are now available for tasks such as the reconstruction of sparsely sampled spectra, peak picking, and virtual decoupling. Here we present a DNN for the analysis of chemical exchange saturation transfer (CEST) data reporting on two- or three-site chemical exchange involving sparse state lifetimes of between approximately 3–60 ms, the range most frequently observed via experiment. The work presented here focuses on the 1H CEST class of methods that are further complicated, in relation to applications to other nuclei, by anti-phase features. The developed DNNs accurately predict the chemical shifts of nuclei in the exchanging species directly from anti-phase 1HN CEST profiles, along with an uncertainty associated with the predictions. The performance of the DNN was quantitatively assessed using both synthetic and experimental anti-phase CEST profiles. The assessments show that the DNN accurately determines chemical shifts and their associated uncertainties. The DNNs developed here do not contain any parameters for the end-user to adjust and the method therefore allows for autonomous analysis of complex NMR data that report on conformational exchange.