Joining dissimilar materials encounters significant engineering challenges due to the contrast in material properties that makes conventional welding not feasible.Magnetic Pulse Welding(MPW)offers a solidstate joining...Joining dissimilar materials encounters significant engineering challenges due to the contrast in material properties that makes conventional welding not feasible.Magnetic Pulse Welding(MPW)offers a solidstate joining technique that overcomes these issues by using impact to create strong bonds without melting the substrate materials.This study investigates the weldability of aluminum alloy Al-5754 with Al-7075 and MARS 380 steel,used in armouring solutions of defense systems,by the use of MPW.In this work,weldability windows are investigated by varying standoff distances between the coating material and its substrate(0.25-4.5 mm)and discharge energies(5-13 kJ)with both O-shape and U-shape inductors.Mechanical strength of the welded joints were assessed through single lap shear tests,identifying optimal welding parameters.Then,the velocity profiles of the flyer plates were measured using heterodyne velocimetry to understand the dynamics of the impact.Then,substructures assembled with the optimal welding conditions were subjected to ballistic testing using 7.62 mm×51 mm NATO and 9 mm×19 mm Parabellum munitions to evaluate the resilience of the welds under ballistic impact.The outcomes demonstrate that MPW effectively joins Al-5754 with both Al-7075 and MARS 380,producing robust welds capable of withstanding ballistic impacts under certain conditions.This research advances the application of MPW in lightweight ballistic protection of defense systems,contributing to the development of more resilient and lighter protective structures.展开更多
We designed and investigated a passive synchronized mode-locked fiber laser.The device utilizes a dual-cavity structure driven by the nonlinear polarization rotation(NPR)mechanism.Stable mode-locking is attained by sy...We designed and investigated a passive synchronized mode-locked fiber laser.The device utilizes a dual-cavity structure driven by the nonlinear polarization rotation(NPR)mechanism.Stable mode-locking is attained by synergistically controlling gain,polarization state,and optical path length in two symmetric sub-cavities.Experiments proved that repetition rate of the sub-cavities can be adjusted via the time delay line(TDL)to achieve synchronized mode-locking.The system stably generates multi-wavelength pulses at a single repetition frequency,evidenced by multiple spectral peaks and equidistant pulse sequences.These findings facilitate the development of high-performance multi-wavelength ultrashort pulse sources,crucial for optical communications,spectral analysis,and remote sensing.展开更多
Low-density short-duration pulsed current-assisted aging treatment was applied to the Ti-6Al-4V-0.5Mo-0.5Zr alloy subjected to different solution treatments.The results show that numerous α_(p) phases redissolve into...Low-density short-duration pulsed current-assisted aging treatment was applied to the Ti-6Al-4V-0.5Mo-0.5Zr alloy subjected to different solution treatments.The results show that numerous α_(p) phases redissolve into the new β phase during the pulsed current-assisted aging process,and then the newly formed β phase is mainly transformed into the β_(t) phase,with occasional transition to new α_(p) phase,leading to a remarkable grain refinement,especially for the lamellarαs phases.In comparison to conventional aging treatment,the pulsed current-assisted aging approach achieves a significant enhancement in strength without degrading ductility,yielding an excellent mechanical property combination:a yield strength of 932 MPa,a tensile strength of 1042 MPa,and an elongation of 12.2%.It is primarily ascribed to the increased fraction of β_(t) phases,the obvious grain refinement effect,and the slip block effect induced by the multiple-variantαs colonies distributed within β_(t) phases.展开更多
Dissimilar AZ31B magnesium alloy and DC56D steel were welded via AA1060 aluminum alloy by magnetic pulse welding.The effects of primary and secondary welding processes on the welded interface were comparatively invest...Dissimilar AZ31B magnesium alloy and DC56D steel were welded via AA1060 aluminum alloy by magnetic pulse welding.The effects of primary and secondary welding processes on the welded interface were comparatively investigated.Macroscopic morphology,microstructure,and interfacial structure of the joints were analyzed using scanning electron microscope,energy dispersive spectrometer,and X-ray diffractometer(XRD).The results show that magnetic pulse welding of dissimilar Mg/Fe metals is achieved using an Al interlayer,which acts as a bridge for deformation and diffusion.Specifically,the AZ31B/AA1060 interface exhibits a typical wavy morphology,and a transition zone exists at the joint interface,which may result in an extremely complex microstructure.The microstructure of this transition zone differs from that of AZ31B magnesium and 1060 Al alloys,and it is identified as brittle intermetallic compounds(IMCs)Al_(3)Mg_(2) and Al_(12)Mg_(17).The transition zone is mainly distributed on the Al side,with the maximum thickness of Al-side transition layer reaching approximately 13.53μm.Incomplete melting layers with varying thicknesses are observed at the primary weld interface,while micron-sized hole defects appear in the transition zone of the secondary weld interface.The AA1060/DC56D interface is mainly straight,with only a small number of discontinuous transition zones distributed intermittently along the interface.These transition zones are characterized by the presence of the brittle IMC FeAl_(3),with a maximum thickness of about 4μm.展开更多
Cold deformation treatment of Cu-bearing stainless steel through a cold rolling process combined with electric pulse treatment(EPT)can significantly improve the microstructure and formability of cold-rolled Cu-bearing...Cold deformation treatment of Cu-bearing stainless steel through a cold rolling process combined with electric pulse treatment(EPT)can significantly improve the microstructure and formability of cold-rolled Cu-bearing stainless steel.The microstructure after EPT was characterized by scanning electron microscopy,transmission electron microscopy,and in-situ tensile testing.It is found that compared with conventional heat treatment,EPT process can significantly promote the nucleation rate and mobility at grain boundaries of the deformed samples,greatly accelerating the recovery and static recrystallization of Cu-bearing stainless steel samples at lower temperatures and contributing to the recovery of anisotropy and the re-molding of deformed samples.Microstructural characterization and theoretical analyses show that the rapid recrystallization during EPT process is caused not only by Joule heating effects but also by non-thermal effects that accelerate grain boundary migration and dislocation destruction and regeneration.In addition,EPT process significantly accelerated the nucleation and precipitation growth of Cu-rich phase.The coarsening of Cu-rich phase during EPT process is due to not only the high vacancy diffusion coefficient under the action of the electric pulse but also the increase in the vacancy diffusion flux induced by the electromigration in the process of the electric pulse.展开更多
Test case prioritization and ranking play a crucial role in software testing by improving fault detection efficiency and ensuring software reliability.While prioritization selects the most relevant test cases for opti...Test case prioritization and ranking play a crucial role in software testing by improving fault detection efficiency and ensuring software reliability.While prioritization selects the most relevant test cases for optimal coverage,ranking further refines their execution order to detect critical faults earlier.This study investigates machine learning techniques to enhance both prioritization and ranking,contributing to more effective and efficient testing processes.We first employ advanced feature engineering alongside ensemble models,including Gradient Boosted,Support Vector Machines,Random Forests,and Naive Bayes classifiers to optimize test case prioritization,achieving an accuracy score of 0.98847 and significantly improving the Average Percentage of Fault Detection(APFD).Subsequently,we introduce a deep Q-learning framework combined with a Genetic Algorithm(GA)to refine test case ranking within priority levels.This approach achieves a rank accuracy of 0.9172,demonstrating robust performance despite the increasing computational demands of specialized variation operators.Our findings highlight the effectiveness of stacked ensemble learning and reinforcement learning in optimizing test case prioritization and ranking.This integrated approach improves testing efficiency,reduces late-stage defects,and improves overall software stability.The study provides valuable information for AI-driven testing frameworks,paving the way for more intelligent and adaptive software quality assurance methodologies.展开更多
With the rapid development of artificial intelligence,the intelligence level of software is increasingly improving.Intelligent software,which is widely applied in crucial fields such as autonomous driving,intelligent ...With the rapid development of artificial intelligence,the intelligence level of software is increasingly improving.Intelligent software,which is widely applied in crucial fields such as autonomous driving,intelligent customer service,and medical diagnosis,is constructed based on complex technologies like machine learning and deep learning.Its uncertain behavior and data dependence pose unprecedented challenges to software testing.However,existing software testing courses mainly focus on conventional contents and are unable to meet the requirements of intelligent software testing.Therefore,this work deeply analyzed the relevant technologies of intelligent software testing,including reliability evaluation indicator system,neuron coverage,and test case generation.It also systematically designed an intelligent software testing course,covering teaching objectives,teaching content,teaching methods,and a teaching case.Verified by the practical teaching in four classes,this course has achieved remarkable results,providing practical experience for the reform of software testing courses.展开更多
Members of the British Textile Machinery Association(BTMA)can look back on 2025 as a year marked by notable technological advances and continued progress in global trade,despite an uncertain and volatile market.“Our ...Members of the British Textile Machinery Association(BTMA)can look back on 2025 as a year marked by notable technological advances and continued progress in global trade,despite an uncertain and volatile market.“Our members have been very active over the past 12 months and this has resulted in new technologies for the production of technical fibres and fabrics,the introduction of AI and machine learning into process control systems and significant advances in materials testing,”says BTMA CEO Jason Kent.“There’s real excitement about what can be achieved in 2026 as we look ahead to upcoming exhibitions such as JEC Composites in Paris in March and Techtextil in Frankfurt in April.”展开更多
In the past few years,efforts have been made to extend the sensitivity of surface nuclear magnetic resonance(SNMR)to short relaxation times,typical for strongly bound water,which,for example,occurs in partially satura...In the past few years,efforts have been made to extend the sensitivity of surface nuclear magnetic resonance(SNMR)to short relaxation times,typical for strongly bound water,which,for example,occurs in partially saturated soils.The two limiting factors for the sensitivity are the dead time after the excitation pulse and the duration of the pulse itself.To enable short pulses,while also achieving proper depths of investigation,high pulse amplitudes are needed.This makes it necessary to consider the Bloch-Siegert effect,i.e.the counter-rotating component and the parallel component of the excitation field have significant influence on the excitation.If an untuned transmitter circuit is used,the pulse shape will also be non-sinusoidal.In this paper,we demonstrate that this influences SNMR measurements with short pulses in two ways:On one hand,the pulse shape influences the phase of the fundamental frequency oscillation.On the other,at very high pulse amplitudes,other frequency components of the excitation field start to influence the excitation.The behavior of the macroscopic magnetizations in the subsurface during the pulse is simulated by solving the Bloch equations,using the pulse shape as an input.Since these calculations are computational expensive,we propose a lookup scheme that allows a time efficient modeling of the obtained SNMR data.展开更多
Pulse tube cryocoolers are widely employed in cryogenic systems,where gas contamination has become a critical factor limiting both performance and service life.To further investigate the condensation behavior of conta...Pulse tube cryocoolers are widely employed in cryogenic systems,where gas contamination has become a critical factor limiting both performance and service life.To further investigate the condensation behavior of contaminants,this study develops a two-dimensional axisymmetric model of a linear-type cryocooler to simulate the transport and deposition processes of trace CO_(2),evaluating the impact of contamination on system pressure drop under various operating conditions.Results indicate that CO_(2)diffusion is primarily driven by concentration gradients.The CO_(2)deposition rate increases markedly at low temperatures and high concentrations,with over 90%of deposition occurring in the cold-end heat exchanger.Under different concentration distributions,dry ice predominantly accumulates in the cold-end heat exchanger;however,notable differences emerge in the pulse tube.In the uniform distribution case,CO_(2)tends to deposit along the inner wall of the pulse tube,whereas in the gradual release scenario,deposition mainly occurs on the cold-end flow straightening mesh screen.Dry ice deposition significantly increases the pressure drop across the system and decreases the pressure wave amplitude,resulting in a degradation of cooling capacity.This study lays a foundation for further investigation into the thermal properties of contaminant layers and provides theoretical guidance for optimizing cold-end components to improve contamination resilience.展开更多
The susceptibility of ore particles to electrical breakdown plays a critical role for high voltage pulse(HVP)breakage,yet its quantitative characterization still lacks deep understanding.Two indicators,namely breakdow...The susceptibility of ore particles to electrical breakdown plays a critical role for high voltage pulse(HVP)breakage,yet its quantitative characterization still lacks deep understanding.Two indicators,namely breakdown delay time(T_(d))and breakdown strength(E_(b))were compared,based on analysis on the two breakdown modes namely wavefront mode and post-wave mode.It was found that T_(d) is more suitable to characterize the susceptibility of ore particles to electrical breakdown in HVP breakage than E_(b).A probabilistic model based on the Weibull distribution is developed to describe the relation of breakdown probability to T_(d).Regression analyses were conducted to investigate how operating parameters and particle properties influence Td and size reduction degree of ore particles in HVP breakage.The regressed models demonstrate potential capability to predict metallic minerals content and HVP breakage degree based on operating parameters and particle properties.展开更多
High-voltage electric pulse(HVEP)rock fragmentation has demonstrated substantial potential for sustainable fracturing of hard rocks owing to its energy efficiency.The transient nature and highly disruptive characteris...High-voltage electric pulse(HVEP)rock fragmentation has demonstrated substantial potential for sustainable fracturing of hard rocks owing to its energy efficiency.The transient nature and highly disruptive characteristics of its physical fracturing process render experimental investigation of the underlying rock-breaking mechanisms challenging.However,existing numerical studies lack comprehensive models that precisely link electrical breakdown phenomena with mechanical disintegration processes.This study combines COMSOL electrical breakdown simulations with four-dimension lattice spring model(4D-LSM)mechanical analysis to establish a coupled HVEP rock fragmentation model.The core concept of the model construction is to import the temperature field of the plasma channel obtained from the electrical breakdown into the mechanical solver to realize the precise connection between the two stages.The validated numerical model elucidates the full process of HVEP-induced fragmentation under varying electrical parameters.Furthermore,the effects of confining pressure and mineral grain size on fragmentation behavior have been investigated.Finally,parametric simulations across 25 electrical parameter combinations demonstrate the critical role of electrode spacing optimization in achieving energy-efficient rock fragmentation.These findings provide a predictive tool for designing efficient HVEP systems in deep resource extraction and mineral processing engineering.展开更多
With the rapid development of Internet technology,REST APIs(Representational State Transfer Application Programming Interfaces)have become the primary communication standard in modern microservice architectures,raisin...With the rapid development of Internet technology,REST APIs(Representational State Transfer Application Programming Interfaces)have become the primary communication standard in modern microservice architectures,raising increasing concerns about their security.Existing fuzz testing methods include random or dictionary-based input generation,which often fail to ensure both syntactic and semantic correctness,and OpenAPIbased approaches,which offer better accuracy but typically lack detailed descriptions of endpoints,parameters,or data formats.To address these issues,this paper proposes the APIDocX fuzz testing framework.It introduces a crawler tailored for dynamic web pages that automatically simulates user interactions to trigger APIs,capturing and extracting parameter information from communication packets.A multi-endpoint parameter adaptation method based on improved Jaccard similarity is then used to generalize these parameters to other potential API endpoints,filling in gaps in OpenAPI specifications.Experimental results demonstrate that the extracted parameters can be generalized with 79.61%accuracy.Fuzz testing using the enriched OpenAPI documents leads to improvements in test coverage,the number of valid test cases generated,and fault detection capabilities.This approach offers an effective enhancement to automated REST API security testing.展开更多
Cloud services,favored by many enterprises due to their high flexibility and easy operation,are widely used for data storage and processing.However,the high latency,together with transmission overheads of the cloud ar...Cloud services,favored by many enterprises due to their high flexibility and easy operation,are widely used for data storage and processing.However,the high latency,together with transmission overheads of the cloud architecture,makes it difficult to quickly respond to the demands of IoT applications and local computation.To make up for these deficiencies in the cloud,fog computing has emerged as a critical role in the IoT applications.It decentralizes the computing power to various lower nodes close to data sources,so as to achieve the goal of low latency and distributed processing.With the data being frequently exchanged and shared between multiple nodes,it becomes a challenge to authorize data securely and efficiently while protecting user privacy.To address this challenge,proxy re-encryption(PRE)schemes provide a feasible way allowing an intermediary proxy node to re-encrypt ciphertext designated for different authorized data requesters without compromising any plaintext information.Since the proxy is viewed as a semi-trusted party,it should be taken to prevent malicious behaviors and reduce the risk of data leakage when implementing PRE schemes.This paper proposes a new fog-assisted identity-based PRE scheme supporting anonymous key generation,equality test,and user revocation to fulfill various IoT application requirements.Specifically,in a traditional identity-based public key architecture,the key escrow problem and the necessity of a secure channel are major security concerns.We utilize an anonymous key generation technique to solve these problems.The equality test functionality further enables a cloud server to inspect whether two candidate trapdoors contain an identical keyword.In particular,the proposed scheme realizes fine-grained user-level authorization while maintaining strong key confidentiality.To revoke an invalid user identity,we add a revocation list to the system flows to restrict access privileges without increasing additional computation cost.To ensure security,it is shown that our system meets the security notion of IND-PrID-CCA and OW-ID-CCA under the Decisional Bilinear Diffie-Hellman(DBDH)assumption.展开更多
The capture zones of the continuous and pulsed guidance laws in the pursuit-evasion game are analytically discussed in this paper to provide deep insights into the capturability distinction between the continuous guid...The capture zones of the continuous and pulsed guidance laws in the pursuit-evasion game are analytically discussed in this paper to provide deep insights into the capturability distinction between the continuous guidance law and the pulsed guidance law.Specifically,first,in the pursuit-evasion game,various capture cases are defined regarding the Zero-Effort Miss distance(ZEM)to facilitate the capturability analysis.Then,for both the evader and the pursuer,the Linear-Quadratic Differential Game(LQDG)guidance laws concerning the continuous acceleration and the pulsed acceleration are converted into a unified form.In each capture case,the optimal solution existence conditions are derived,and the corresponding capture zones are formulated.The discussion on the capture zones shows that if the optimal solution exists,the distinction between the pulsed guidance law and the continuous guidance law can be neglected under small guidance effort weight.However,the capture zone of the continuous guidance law is larger than that of the pulsed guidance law with large pursuer guidance effort weight,but smaller with large evader guidance effort weight.Finally,various simulations are conducted to illustrate the distinction of the continuous and pulsed guidance laws,as well as the impact of the acceleration ratio and the time constant ratio on the capturability.展开更多
Lateral flow immunoassays(LFIAs)are low-cost,rapid,and easy to use for pointof-care testing(POCT),but the majority of the available LFIA tests are indicative,rather than quantitative,and their sensitivity in antigen t...Lateral flow immunoassays(LFIAs)are low-cost,rapid,and easy to use for pointof-care testing(POCT),but the majority of the available LFIA tests are indicative,rather than quantitative,and their sensitivity in antigen tests are usually limited at the nanogram range,which is primarily due to the passive capillary fluidics through nitrocellulose membranes,often associated with non-specific bindings and high background noise.To overcome this challenge,we report a Beads-on-a-Tip design by replacing nitrocellulose membranes with a pipette tip loaded with magnetic beads.The beads are pre-conjugated with capture antibodies that support a typical sandwich immunoassay.This design enriches the low-abundant antigen proteins and allows an active washing process to significantly reduce non-specific bindings.To further improve the detection sensitivity,we employed upconversion nanoparticles(UCNPs)as luminescent reporters and SARS-CoV-2 spike(S)antigen as a model analyte to benchmark the performance of this design against our previously reported methods.We found that the key to enhance the immunocomplex formation and signal-to-noise ratio lay in optimizing incubation time and the UCNP-to-bead ratio.We therefore successfully demonstrated that the new method can achieve a very large dynamic range from 500 fg/mL to 10μg/mL,across over 7 digits,and a limit of detection of 706 fg/mL,nearly another order of magnitude lower than the best reported LFIA using UCNPs in COVID-19 spike antigen detection.Our system offers a promising solution for ultra-sensitive and quantitative POCT diagnostics.展开更多
Pulsed dynamic electrolysis(PDE),driven by renewable energy,has emerged as an innovative electrocatalytic conversion method,demonstrating significant potential in addressing global energy challenges and promoting sust...Pulsed dynamic electrolysis(PDE),driven by renewable energy,has emerged as an innovative electrocatalytic conversion method,demonstrating significant potential in addressing global energy challenges and promoting sustainable development.Despite significant progress in various electrochemical systems,the regulatory mechanisms of PDE in energy and mass transfer and the lifespan extension of electrolysis systems,particularly in water electrolysis(WE)for hydrogen production,remain insufficiently explored.Therefore,there is an urgent need for a deeper understanding of the unique contributions of PDE in mass transfer enhancement,microenvironment regulation,and hydrogen production optimization,aiming to achieve low-energy consumption,high catalytic activity,and long-term stability in the generation of target products.Here,this review critically examines the microenvironmental effects of PDE on energy and mass transfer,the electrode degradation mechanisms in the lifespan extension of electrolysis systems,and the key factors in enhancing WE for hydrogen production,providing a comprehensive summary of current research progress.The review focuses on the complex regulatory mechanisms of frequency,duty cycle,amplitude,and other factors in hydrogen evolution reaction(HER)performance within PDE strategies,revealing the interrelationships among them.Finally,the potential future directions and challenges for transitioning from laboratory studies to industrial applications are proposed.展开更多
To address the insufficient prediction accuracy of multi-state parameters in electro-hydraulic servo material fatigue testing machines under complex loading and nonlinear coupling conditions,this paper proposes a mult...To address the insufficient prediction accuracy of multi-state parameters in electro-hydraulic servo material fatigue testing machines under complex loading and nonlinear coupling conditions,this paper proposes a multivariate sequence-to-sequence prediction model integrating a Long Short-Term Memory(LSTM)encoder,a Gated Recurrent Unit(GRU)decoder,and a multi-head attention mechanism.This approach enhances prediction accuracy and robustness across different control modes and load spectra by leveraging multi-channel inputs and cross-variable feature interactions,thereby capturing both short-term high-frequency dynamics and long-term slow drift characteristics.Experiments using long-term data from real test benches demonstrate that the model achieves a stable MSE below 0.01 on the validation set,with MAE and RMSE of approximately 0.018 and 0.052,respectively,and a coefficient of determination reaching 0.98.This significantly outperforms traditional identification methods and single RNN models.Sensitivity analysis indicates that a prediction stride of 10 achieves an optimal balance between accuracy and computational overhead.Ablation experiments validated the contribution of multi-head attention and decoder architecture to enhancing cross-variable coupling modeling capabilities.This model can be applied to residualdriven early warning in health monitoring,and risk assessment with scheme optimization in test design.It enables near-real-time deployment feasibility,providing a practical data-driven technical pathway for reliability assurance in advanced equipment.展开更多
Patients affected by monogenic diseases impose a substantial burden on both themselves and their families.The primary preventive measure,i.e.,invasive prenatal diagnosis,carries a risk of miscarriage and cannot be per...Patients affected by monogenic diseases impose a substantial burden on both themselves and their families.The primary preventive measure,i.e.,invasive prenatal diagnosis,carries a risk of miscarriage and cannot be performed early in pregnancy.Hence,there is a need for non-invasive prenatal testing(NIPT)for monogenic diseases.By utilizing enriched cell-free fetal DNA(cffDNA)from maternal plasma,we refine the NIPT method,which combines targeted region capture technology,haplotyping,and analysis of informative site frequency.We apply this method to 93 clinical families at genetic risk for thalassemia,encompassing various genetic variant types,to establish a workflow and evaluate its efficiency.Our approach requires only 3 ng of DNA input to generate 0.1 Gb informative target genomic data and leverages a minimum of 3%cffDNA.This method has a 98.16%success rate and 100%concordance with conventional invasive methods.Furthermore,we demonstrate the ability to analyze fetal genotypes as early as eight weeks of gestation.This study establishes an optimized NIPT method for the early detection of various thalassemia disorders during pregnancy.This technique demonstrates high accuracy and potential for clinical application in prenatal diagnosis.展开更多
As deep learning(DL)models are increasingly deployed in sensitive domains(e.g.,healthcare),concerns over privacy and security have intensified.Conventional penetration testing frameworks,such asOWASP and NIST,are effe...As deep learning(DL)models are increasingly deployed in sensitive domains(e.g.,healthcare),concerns over privacy and security have intensified.Conventional penetration testing frameworks,such asOWASP and NIST,are effective for traditional networks and applications but lack the capabilities to address DL-specific threats,such asmodel inversion,membership inference,and adversarial attacks.This review provides a comprehensive analysis of penetration testing for the privacy of DL models,examining the shortfalls of existing frameworks,tools,and testing methodologies.Through systematic evaluation of existing literature and empirical analysis,we identify three major contributions:(i)a critical assessment of traditional penetration testing frameworks’inadequacies when applied to DL-specific privacy vulnerabilities,(ii)a comprehensive evaluation of state-of-the-art privacy-preserving methods and their integration with penetration testing workflows,and(iii)the development of a structured framework that combines reconnaissance,threat modeling,exploitation,and post-exploitation phases specifically tailored for DL privacy assessment.Moreover,this review evaluates popular solutions such as IBMAdversarial Robustness Toolbox and TensorFlowPrivacy,alongside privacy-preserving techniques(e.g.,Differential Privacy,Homomorphic Encryption,and Federated Learning),which we systematically analyze through comparative studies of their effectiveness,computational overhead,and practical deployment constraints.While these techniques offer promising safeguards,their adoption is hindered by accuracy loss,performance overheads,and the rapid evolution of attack strategies.Our findings reveal that no single existing solution provides comprehensive protection,which leads us to propose a hybrid approach that strategically combines multiple privacy-preserving mechanisms.The findings of this survey underscore an urgent need for automated,regulationcompliant penetration testing frameworks specifically tailored to DL systems.We argue for hybrid privacy solutions that combinemultiple protectivemechanisms to ensure bothmodel accuracy and privacy.Building on our analysis,we present actionable recommendations for developing adaptive penetration testing strategies that incorporate automated vulnerability assessment,continuous monitoring,and regulatory compliance verification.展开更多
基金funded on the one hand by Agence de l'Innovation de Défense(AID)grant reference number 2021650044on the other hand by Ecole Centrale de Nantes。
文摘Joining dissimilar materials encounters significant engineering challenges due to the contrast in material properties that makes conventional welding not feasible.Magnetic Pulse Welding(MPW)offers a solidstate joining technique that overcomes these issues by using impact to create strong bonds without melting the substrate materials.This study investigates the weldability of aluminum alloy Al-5754 with Al-7075 and MARS 380 steel,used in armouring solutions of defense systems,by the use of MPW.In this work,weldability windows are investigated by varying standoff distances between the coating material and its substrate(0.25-4.5 mm)and discharge energies(5-13 kJ)with both O-shape and U-shape inductors.Mechanical strength of the welded joints were assessed through single lap shear tests,identifying optimal welding parameters.Then,the velocity profiles of the flyer plates were measured using heterodyne velocimetry to understand the dynamics of the impact.Then,substructures assembled with the optimal welding conditions were subjected to ballistic testing using 7.62 mm×51 mm NATO and 9 mm×19 mm Parabellum munitions to evaluate the resilience of the welds under ballistic impact.The outcomes demonstrate that MPW effectively joins Al-5754 with both Al-7075 and MARS 380,producing robust welds capable of withstanding ballistic impacts under certain conditions.This research advances the application of MPW in lightweight ballistic protection of defense systems,contributing to the development of more resilient and lighter protective structures.
文摘We designed and investigated a passive synchronized mode-locked fiber laser.The device utilizes a dual-cavity structure driven by the nonlinear polarization rotation(NPR)mechanism.Stable mode-locking is attained by synergistically controlling gain,polarization state,and optical path length in two symmetric sub-cavities.Experiments proved that repetition rate of the sub-cavities can be adjusted via the time delay line(TDL)to achieve synchronized mode-locking.The system stably generates multi-wavelength pulses at a single repetition frequency,evidenced by multiple spectral peaks and equidistant pulse sequences.These findings facilitate the development of high-performance multi-wavelength ultrashort pulse sources,crucial for optical communications,spectral analysis,and remote sensing.
基金National Key Research and Development Program of China(2021YFB3700801)。
文摘Low-density short-duration pulsed current-assisted aging treatment was applied to the Ti-6Al-4V-0.5Mo-0.5Zr alloy subjected to different solution treatments.The results show that numerous α_(p) phases redissolve into the new β phase during the pulsed current-assisted aging process,and then the newly formed β phase is mainly transformed into the β_(t) phase,with occasional transition to new α_(p) phase,leading to a remarkable grain refinement,especially for the lamellarαs phases.In comparison to conventional aging treatment,the pulsed current-assisted aging approach achieves a significant enhancement in strength without degrading ductility,yielding an excellent mechanical property combination:a yield strength of 932 MPa,a tensile strength of 1042 MPa,and an elongation of 12.2%.It is primarily ascribed to the increased fraction of β_(t) phases,the obvious grain refinement effect,and the slip block effect induced by the multiple-variantαs colonies distributed within β_(t) phases.
文摘Dissimilar AZ31B magnesium alloy and DC56D steel were welded via AA1060 aluminum alloy by magnetic pulse welding.The effects of primary and secondary welding processes on the welded interface were comparatively investigated.Macroscopic morphology,microstructure,and interfacial structure of the joints were analyzed using scanning electron microscope,energy dispersive spectrometer,and X-ray diffractometer(XRD).The results show that magnetic pulse welding of dissimilar Mg/Fe metals is achieved using an Al interlayer,which acts as a bridge for deformation and diffusion.Specifically,the AZ31B/AA1060 interface exhibits a typical wavy morphology,and a transition zone exists at the joint interface,which may result in an extremely complex microstructure.The microstructure of this transition zone differs from that of AZ31B magnesium and 1060 Al alloys,and it is identified as brittle intermetallic compounds(IMCs)Al_(3)Mg_(2) and Al_(12)Mg_(17).The transition zone is mainly distributed on the Al side,with the maximum thickness of Al-side transition layer reaching approximately 13.53μm.Incomplete melting layers with varying thicknesses are observed at the primary weld interface,while micron-sized hole defects appear in the transition zone of the secondary weld interface.The AA1060/DC56D interface is mainly straight,with only a small number of discontinuous transition zones distributed intermittently along the interface.These transition zones are characterized by the presence of the brittle IMC FeAl_(3),with a maximum thickness of about 4μm.
基金supported by the National Natural Science Foundation of China(Nos.52305401 and 52475391)National Key Research and Development Program of China(2024YFB3714301)+1 种基金the Fundamental Research Program of Shanxi Province(202303011211004,TZLH20230818001)Engineering Research Center of the Ministry of Education.
文摘Cold deformation treatment of Cu-bearing stainless steel through a cold rolling process combined with electric pulse treatment(EPT)can significantly improve the microstructure and formability of cold-rolled Cu-bearing stainless steel.The microstructure after EPT was characterized by scanning electron microscopy,transmission electron microscopy,and in-situ tensile testing.It is found that compared with conventional heat treatment,EPT process can significantly promote the nucleation rate and mobility at grain boundaries of the deformed samples,greatly accelerating the recovery and static recrystallization of Cu-bearing stainless steel samples at lower temperatures and contributing to the recovery of anisotropy and the re-molding of deformed samples.Microstructural characterization and theoretical analyses show that the rapid recrystallization during EPT process is caused not only by Joule heating effects but also by non-thermal effects that accelerate grain boundary migration and dislocation destruction and regeneration.In addition,EPT process significantly accelerated the nucleation and precipitation growth of Cu-rich phase.The coarsening of Cu-rich phase during EPT process is due to not only the high vacancy diffusion coefficient under the action of the electric pulse but also the increase in the vacancy diffusion flux induced by the electromigration in the process of the electric pulse.
文摘Test case prioritization and ranking play a crucial role in software testing by improving fault detection efficiency and ensuring software reliability.While prioritization selects the most relevant test cases for optimal coverage,ranking further refines their execution order to detect critical faults earlier.This study investigates machine learning techniques to enhance both prioritization and ranking,contributing to more effective and efficient testing processes.We first employ advanced feature engineering alongside ensemble models,including Gradient Boosted,Support Vector Machines,Random Forests,and Naive Bayes classifiers to optimize test case prioritization,achieving an accuracy score of 0.98847 and significantly improving the Average Percentage of Fault Detection(APFD).Subsequently,we introduce a deep Q-learning framework combined with a Genetic Algorithm(GA)to refine test case ranking within priority levels.This approach achieves a rank accuracy of 0.9172,demonstrating robust performance despite the increasing computational demands of specialized variation operators.Our findings highlight the effectiveness of stacked ensemble learning and reinforcement learning in optimizing test case prioritization and ranking.This integrated approach improves testing efficiency,reduces late-stage defects,and improves overall software stability.The study provides valuable information for AI-driven testing frameworks,paving the way for more intelligent and adaptive software quality assurance methodologies.
基金Computer Basic Education Teaching Research Project of Association of Fundamental Computing Education in Chinese Universities(Nos.2025-AFCEC-527 and 2024-AFCEC-088)Research on the Reform of Public Course Teaching at Nantong College of Science and Technology(No.2024JGG015).
文摘With the rapid development of artificial intelligence,the intelligence level of software is increasingly improving.Intelligent software,which is widely applied in crucial fields such as autonomous driving,intelligent customer service,and medical diagnosis,is constructed based on complex technologies like machine learning and deep learning.Its uncertain behavior and data dependence pose unprecedented challenges to software testing.However,existing software testing courses mainly focus on conventional contents and are unable to meet the requirements of intelligent software testing.Therefore,this work deeply analyzed the relevant technologies of intelligent software testing,including reliability evaluation indicator system,neuron coverage,and test case generation.It also systematically designed an intelligent software testing course,covering teaching objectives,teaching content,teaching methods,and a teaching case.Verified by the practical teaching in four classes,this course has achieved remarkable results,providing practical experience for the reform of software testing courses.
文摘Members of the British Textile Machinery Association(BTMA)can look back on 2025 as a year marked by notable technological advances and continued progress in global trade,despite an uncertain and volatile market.“Our members have been very active over the past 12 months and this has resulted in new technologies for the production of technical fibres and fabrics,the introduction of AI and machine learning into process control systems and significant advances in materials testing,”says BTMA CEO Jason Kent.“There’s real excitement about what can be achieved in 2026 as we look ahead to upcoming exhibitions such as JEC Composites in Paris in March and Techtextil in Frankfurt in April.”
基金funded by the German Research Foundation(Deutsche Forschungsgemeinschaft-DFG)under grant MU 3318/8-1.
文摘In the past few years,efforts have been made to extend the sensitivity of surface nuclear magnetic resonance(SNMR)to short relaxation times,typical for strongly bound water,which,for example,occurs in partially saturated soils.The two limiting factors for the sensitivity are the dead time after the excitation pulse and the duration of the pulse itself.To enable short pulses,while also achieving proper depths of investigation,high pulse amplitudes are needed.This makes it necessary to consider the Bloch-Siegert effect,i.e.the counter-rotating component and the parallel component of the excitation field have significant influence on the excitation.If an untuned transmitter circuit is used,the pulse shape will also be non-sinusoidal.In this paper,we demonstrate that this influences SNMR measurements with short pulses in two ways:On one hand,the pulse shape influences the phase of the fundamental frequency oscillation.On the other,at very high pulse amplitudes,other frequency components of the excitation field start to influence the excitation.The behavior of the macroscopic magnetizations in the subsurface during the pulse is simulated by solving the Bloch equations,using the pulse shape as an input.Since these calculations are computational expensive,we propose a lookup scheme that allows a time efficient modeling of the obtained SNMR data.
基金supported by the National Natural Science Foundation of China(No.52376012)the Aeronautical Science Foundation of China(20230024047001).
文摘Pulse tube cryocoolers are widely employed in cryogenic systems,where gas contamination has become a critical factor limiting both performance and service life.To further investigate the condensation behavior of contaminants,this study develops a two-dimensional axisymmetric model of a linear-type cryocooler to simulate the transport and deposition processes of trace CO_(2),evaluating the impact of contamination on system pressure drop under various operating conditions.Results indicate that CO_(2)diffusion is primarily driven by concentration gradients.The CO_(2)deposition rate increases markedly at low temperatures and high concentrations,with over 90%of deposition occurring in the cold-end heat exchanger.Under different concentration distributions,dry ice predominantly accumulates in the cold-end heat exchanger;however,notable differences emerge in the pulse tube.In the uniform distribution case,CO_(2)tends to deposit along the inner wall of the pulse tube,whereas in the gradual release scenario,deposition mainly occurs on the cold-end flow straightening mesh screen.Dry ice deposition significantly increases the pressure drop across the system and decreases the pressure wave amplitude,resulting in a degradation of cooling capacity.This study lays a foundation for further investigation into the thermal properties of contaminant layers and provides theoretical guidance for optimizing cold-end components to improve contamination resilience.
基金The financial supports from National Natural Science Foundation of China(Nos.52574313,52204272 and 52074091)to this project。
文摘The susceptibility of ore particles to electrical breakdown plays a critical role for high voltage pulse(HVP)breakage,yet its quantitative characterization still lacks deep understanding.Two indicators,namely breakdown delay time(T_(d))and breakdown strength(E_(b))were compared,based on analysis on the two breakdown modes namely wavefront mode and post-wave mode.It was found that T_(d) is more suitable to characterize the susceptibility of ore particles to electrical breakdown in HVP breakage than E_(b).A probabilistic model based on the Weibull distribution is developed to describe the relation of breakdown probability to T_(d).Regression analyses were conducted to investigate how operating parameters and particle properties influence Td and size reduction degree of ore particles in HVP breakage.The regressed models demonstrate potential capability to predict metallic minerals content and HVP breakage degree based on operating parameters and particle properties.
基金financial support from the National Natural Science Foundation of China(Nos.52209144 and 12472405).
文摘High-voltage electric pulse(HVEP)rock fragmentation has demonstrated substantial potential for sustainable fracturing of hard rocks owing to its energy efficiency.The transient nature and highly disruptive characteristics of its physical fracturing process render experimental investigation of the underlying rock-breaking mechanisms challenging.However,existing numerical studies lack comprehensive models that precisely link electrical breakdown phenomena with mechanical disintegration processes.This study combines COMSOL electrical breakdown simulations with four-dimension lattice spring model(4D-LSM)mechanical analysis to establish a coupled HVEP rock fragmentation model.The core concept of the model construction is to import the temperature field of the plasma channel obtained from the electrical breakdown into the mechanical solver to realize the precise connection between the two stages.The validated numerical model elucidates the full process of HVEP-induced fragmentation under varying electrical parameters.Furthermore,the effects of confining pressure and mineral grain size on fragmentation behavior have been investigated.Finally,parametric simulations across 25 electrical parameter combinations demonstrate the critical role of electrode spacing optimization in achieving energy-efficient rock fragmentation.These findings provide a predictive tool for designing efficient HVEP systems in deep resource extraction and mineral processing engineering.
基金supported by the Open Foundation of Key Laboratory of Cyberspace Security,Ministry of Education of China(KLCS20240211)。
文摘With the rapid development of Internet technology,REST APIs(Representational State Transfer Application Programming Interfaces)have become the primary communication standard in modern microservice architectures,raising increasing concerns about their security.Existing fuzz testing methods include random or dictionary-based input generation,which often fail to ensure both syntactic and semantic correctness,and OpenAPIbased approaches,which offer better accuracy but typically lack detailed descriptions of endpoints,parameters,or data formats.To address these issues,this paper proposes the APIDocX fuzz testing framework.It introduces a crawler tailored for dynamic web pages that automatically simulates user interactions to trigger APIs,capturing and extracting parameter information from communication packets.A multi-endpoint parameter adaptation method based on improved Jaccard similarity is then used to generalize these parameters to other potential API endpoints,filling in gaps in OpenAPI specifications.Experimental results demonstrate that the extracted parameters can be generalized with 79.61%accuracy.Fuzz testing using the enriched OpenAPI documents leads to improvements in test coverage,the number of valid test cases generated,and fault detection capabilities.This approach offers an effective enhancement to automated REST API security testing.
基金supported in part by the National Science and Technology Council of Taiwan under the contract numbers NSTC 114-2221-E-019-055-MY2 and NSTC 114-2221-E-019-069.
文摘Cloud services,favored by many enterprises due to their high flexibility and easy operation,are widely used for data storage and processing.However,the high latency,together with transmission overheads of the cloud architecture,makes it difficult to quickly respond to the demands of IoT applications and local computation.To make up for these deficiencies in the cloud,fog computing has emerged as a critical role in the IoT applications.It decentralizes the computing power to various lower nodes close to data sources,so as to achieve the goal of low latency and distributed processing.With the data being frequently exchanged and shared between multiple nodes,it becomes a challenge to authorize data securely and efficiently while protecting user privacy.To address this challenge,proxy re-encryption(PRE)schemes provide a feasible way allowing an intermediary proxy node to re-encrypt ciphertext designated for different authorized data requesters without compromising any plaintext information.Since the proxy is viewed as a semi-trusted party,it should be taken to prevent malicious behaviors and reduce the risk of data leakage when implementing PRE schemes.This paper proposes a new fog-assisted identity-based PRE scheme supporting anonymous key generation,equality test,and user revocation to fulfill various IoT application requirements.Specifically,in a traditional identity-based public key architecture,the key escrow problem and the necessity of a secure channel are major security concerns.We utilize an anonymous key generation technique to solve these problems.The equality test functionality further enables a cloud server to inspect whether two candidate trapdoors contain an identical keyword.In particular,the proposed scheme realizes fine-grained user-level authorization while maintaining strong key confidentiality.To revoke an invalid user identity,we add a revocation list to the system flows to restrict access privileges without increasing additional computation cost.To ensure security,it is shown that our system meets the security notion of IND-PrID-CCA and OW-ID-CCA under the Decisional Bilinear Diffie-Hellman(DBDH)assumption.
基金co-supported by the National Natural Science Foundation of China(Nos.U24B20157,62203031)the Natural Science Foundation of Beijing Municipality,China(No.4242041)+2 种基金the Natural Science Foundation of Zhejiang Province,China(No.LY24F030002)the Aeronautical ScienceFoundation of China(No.2024Z066051001)the Fundamental Research Funds for the Central Universities of China。
文摘The capture zones of the continuous and pulsed guidance laws in the pursuit-evasion game are analytically discussed in this paper to provide deep insights into the capturability distinction between the continuous guidance law and the pulsed guidance law.Specifically,first,in the pursuit-evasion game,various capture cases are defined regarding the Zero-Effort Miss distance(ZEM)to facilitate the capturability analysis.Then,for both the evader and the pursuer,the Linear-Quadratic Differential Game(LQDG)guidance laws concerning the continuous acceleration and the pulsed acceleration are converted into a unified form.In each capture case,the optimal solution existence conditions are derived,and the corresponding capture zones are formulated.The discussion on the capture zones shows that if the optimal solution exists,the distinction between the pulsed guidance law and the continuous guidance law can be neglected under small guidance effort weight.However,the capture zone of the continuous guidance law is larger than that of the pulsed guidance law with large pursuer guidance effort weight,but smaller with large evader guidance effort weight.Finally,various simulations are conducted to illustrate the distinction of the continuous and pulsed guidance laws,as well as the impact of the acceleration ratio and the time constant ratio on the capturability.
基金financially supported by ARC Linkage project(LP210200642)ARC Center of Excellence for Quantum Biotechnology(grant no.CE230100021)+1 种基金National Health and Medical Research Council Investigator Fellowship—(grant no.APP2017499)Chan Zuckerberg Initiative Deep Tissue Imaging Phase 2(grant no.DT12-0000000182).
文摘Lateral flow immunoassays(LFIAs)are low-cost,rapid,and easy to use for pointof-care testing(POCT),but the majority of the available LFIA tests are indicative,rather than quantitative,and their sensitivity in antigen tests are usually limited at the nanogram range,which is primarily due to the passive capillary fluidics through nitrocellulose membranes,often associated with non-specific bindings and high background noise.To overcome this challenge,we report a Beads-on-a-Tip design by replacing nitrocellulose membranes with a pipette tip loaded with magnetic beads.The beads are pre-conjugated with capture antibodies that support a typical sandwich immunoassay.This design enriches the low-abundant antigen proteins and allows an active washing process to significantly reduce non-specific bindings.To further improve the detection sensitivity,we employed upconversion nanoparticles(UCNPs)as luminescent reporters and SARS-CoV-2 spike(S)antigen as a model analyte to benchmark the performance of this design against our previously reported methods.We found that the key to enhance the immunocomplex formation and signal-to-noise ratio lay in optimizing incubation time and the UCNP-to-bead ratio.We therefore successfully demonstrated that the new method can achieve a very large dynamic range from 500 fg/mL to 10μg/mL,across over 7 digits,and a limit of detection of 706 fg/mL,nearly another order of magnitude lower than the best reported LFIA using UCNPs in COVID-19 spike antigen detection.Our system offers a promising solution for ultra-sensitive and quantitative POCT diagnostics.
基金financially supported by the Key Research and Development Program of Heilongjiang Province(No.2024ZXJ03C06)National Natural Science Foundation of China(No.52476192,No.52106237)+1 种基金Natural Science Foundation of Heilongjiang Province(No.YQ2022E027)Technology Project of China Datang Technology Innovation Co.,Ltd(No.DTKC-2024-20610).
文摘Pulsed dynamic electrolysis(PDE),driven by renewable energy,has emerged as an innovative electrocatalytic conversion method,demonstrating significant potential in addressing global energy challenges and promoting sustainable development.Despite significant progress in various electrochemical systems,the regulatory mechanisms of PDE in energy and mass transfer and the lifespan extension of electrolysis systems,particularly in water electrolysis(WE)for hydrogen production,remain insufficiently explored.Therefore,there is an urgent need for a deeper understanding of the unique contributions of PDE in mass transfer enhancement,microenvironment regulation,and hydrogen production optimization,aiming to achieve low-energy consumption,high catalytic activity,and long-term stability in the generation of target products.Here,this review critically examines the microenvironmental effects of PDE on energy and mass transfer,the electrode degradation mechanisms in the lifespan extension of electrolysis systems,and the key factors in enhancing WE for hydrogen production,providing a comprehensive summary of current research progress.The review focuses on the complex regulatory mechanisms of frequency,duty cycle,amplitude,and other factors in hydrogen evolution reaction(HER)performance within PDE strategies,revealing the interrelationships among them.Finally,the potential future directions and challenges for transitioning from laboratory studies to industrial applications are proposed.
基金supported by Natural Science Foundation of China(NSFC),Grant number 5247052693.
文摘To address the insufficient prediction accuracy of multi-state parameters in electro-hydraulic servo material fatigue testing machines under complex loading and nonlinear coupling conditions,this paper proposes a multivariate sequence-to-sequence prediction model integrating a Long Short-Term Memory(LSTM)encoder,a Gated Recurrent Unit(GRU)decoder,and a multi-head attention mechanism.This approach enhances prediction accuracy and robustness across different control modes and load spectra by leveraging multi-channel inputs and cross-variable feature interactions,thereby capturing both short-term high-frequency dynamics and long-term slow drift characteristics.Experiments using long-term data from real test benches demonstrate that the model achieves a stable MSE below 0.01 on the validation set,with MAE and RMSE of approximately 0.018 and 0.052,respectively,and a coefficient of determination reaching 0.98.This significantly outperforms traditional identification methods and single RNN models.Sensitivity analysis indicates that a prediction stride of 10 achieves an optimal balance between accuracy and computational overhead.Ablation experiments validated the contribution of multi-head attention and decoder architecture to enhancing cross-variable coupling modeling capabilities.This model can be applied to residualdriven early warning in health monitoring,and risk assessment with scheme optimization in test design.It enables near-real-time deployment feasibility,providing a practical data-driven technical pathway for reliability assurance in advanced equipment.
基金supported by the National Key R&D Program of China(2024YFA1802300)the Major Science and Technology Program of Hainan Province(ZDKJ2021037)+4 种基金the Regional Innovation and Development Joint Fund of the National Natural Science Foundation of China(U24A20677)Hainan Province Science and Technology Special Fund(ZDYF2020117,ZDY2024SHFZ143)Hainan Province Science and TechnologyProject(LCXY202102,LCYX202203,LCYX202301,LCYx202502)Innovative research project for postgraduate students in Hainan Medical University(HYYB2021A05)the Hainan Province Clinical Medical Center,and the specific research fund of The Innovation Platform for Academicians of Hainan Province(YSPTZX202310).
文摘Patients affected by monogenic diseases impose a substantial burden on both themselves and their families.The primary preventive measure,i.e.,invasive prenatal diagnosis,carries a risk of miscarriage and cannot be performed early in pregnancy.Hence,there is a need for non-invasive prenatal testing(NIPT)for monogenic diseases.By utilizing enriched cell-free fetal DNA(cffDNA)from maternal plasma,we refine the NIPT method,which combines targeted region capture technology,haplotyping,and analysis of informative site frequency.We apply this method to 93 clinical families at genetic risk for thalassemia,encompassing various genetic variant types,to establish a workflow and evaluate its efficiency.Our approach requires only 3 ng of DNA input to generate 0.1 Gb informative target genomic data and leverages a minimum of 3%cffDNA.This method has a 98.16%success rate and 100%concordance with conventional invasive methods.Furthermore,we demonstrate the ability to analyze fetal genotypes as early as eight weeks of gestation.This study establishes an optimized NIPT method for the early detection of various thalassemia disorders during pregnancy.This technique demonstrates high accuracy and potential for clinical application in prenatal diagnosis.
基金supported in part by the Tianjin Natural Science Foundation Project(24JCZDJC01000)the Fundamental Research Funds for the Central Universities of China(No.3122025091).
文摘As deep learning(DL)models are increasingly deployed in sensitive domains(e.g.,healthcare),concerns over privacy and security have intensified.Conventional penetration testing frameworks,such asOWASP and NIST,are effective for traditional networks and applications but lack the capabilities to address DL-specific threats,such asmodel inversion,membership inference,and adversarial attacks.This review provides a comprehensive analysis of penetration testing for the privacy of DL models,examining the shortfalls of existing frameworks,tools,and testing methodologies.Through systematic evaluation of existing literature and empirical analysis,we identify three major contributions:(i)a critical assessment of traditional penetration testing frameworks’inadequacies when applied to DL-specific privacy vulnerabilities,(ii)a comprehensive evaluation of state-of-the-art privacy-preserving methods and their integration with penetration testing workflows,and(iii)the development of a structured framework that combines reconnaissance,threat modeling,exploitation,and post-exploitation phases specifically tailored for DL privacy assessment.Moreover,this review evaluates popular solutions such as IBMAdversarial Robustness Toolbox and TensorFlowPrivacy,alongside privacy-preserving techniques(e.g.,Differential Privacy,Homomorphic Encryption,and Federated Learning),which we systematically analyze through comparative studies of their effectiveness,computational overhead,and practical deployment constraints.While these techniques offer promising safeguards,their adoption is hindered by accuracy loss,performance overheads,and the rapid evolution of attack strategies.Our findings reveal that no single existing solution provides comprehensive protection,which leads us to propose a hybrid approach that strategically combines multiple privacy-preserving mechanisms.The findings of this survey underscore an urgent need for automated,regulationcompliant penetration testing frameworks specifically tailored to DL systems.We argue for hybrid privacy solutions that combinemultiple protectivemechanisms to ensure bothmodel accuracy and privacy.Building on our analysis,we present actionable recommendations for developing adaptive penetration testing strategies that incorporate automated vulnerability assessment,continuous monitoring,and regulatory compliance verification.
