The main challenge in AI governance today is striking a balance between controlling AI dangers and fostering AI innovation.Regulators in a number of nations have progressively extended the regulatory sandbox,which was...The main challenge in AI governance today is striking a balance between controlling AI dangers and fostering AI innovation.Regulators in a number of nations have progressively extended the regulatory sandbox,which was first implemented in the banking sector,to AI governance in an effort to reduce the conflict between regulation and innovation.The AI regulatory sandbox is a new and feasible route for AI governance in China that not only helps to manage the risks of technology application but also prevents inhibiting AI innovation.It keeps inventors'trial-and-error tolerance space inside the regulatory purview while offering a controlled setting for the development and testing of novel AI that hasn't yet been put on the market.By providing full-cycle governance of AI with the principles of agility and inclusive prudence,the regulatory sandbox offers an alternative to the conventional top-down hard regulation,expost regulation,and tight regulation.However,the current system also has inherent limitations and practical obstacles that need to be overcome by a more rational and effective approach.To achieve its positive impact on AI governance,the AI regulatory sandbox system should build and improve the access and exit mechanism,the coordination mechanism between the sandbox and personal information protection,and the mechanisms of exemption,disclosure,and communication.展开更多
Fault formation and evolution in the presence of multiple pre-existing weaknesses has not been investigated extensively in rift basins. The fault systems of Weixinan Sag, Beibuwan Basin of China, which is fully covere...Fault formation and evolution in the presence of multiple pre-existing weaknesses has not been investigated extensively in rift basins. The fault systems of Weixinan Sag, Beibuwan Basin of China, which is fully covered with high-precision 3-D seismic data and is rich in oil-gas resources, have been successfully reproduced by sandbox modeling in this study with inclusion of multiple pre-existing weaknesses in the experimental model. The basic characteristics of fault formation and evolution revealed by sandbox modeling are as follows. 1） Weakness-reactivation faults and weakness-related faults are formed much earlier than the distant-weakness faults （faults far away from and with little or no relationship to the weakness）. 2） Weakness-reactivation faults and weakness-related faults develop mainly along or parallel to a pre-existing weakness, while distant-weakness faults develop nearly perpendicular to the extension direction. A complicated fault system can be formed in a fixed direction of extension with the existence of multiple pre-existing weaknesses, and the complicated fault system in the Weixinan Sag formed gradually in a nearly N-S direction with multiple pre-existing weaknesses. 3） The increase in the length and number of faults is closely tied to the nature of pre-existing weaknesses. The sandbox model may provide a new clue to detailed fault system research for oil and gas exploration in rift basins.展开更多
Three series of sandbox modeling experiments were performed to study the fault-increment pattern in extensional basins. Experimental results showed that the tectonic action mode of boundaries and the shape of major bo...Three series of sandbox modeling experiments were performed to study the fault-increment pattern in extensional basins. Experimental results showed that the tectonic action mode of boundaries and the shape of major boundary faults control the formation and evolution of faults in extensional basins. In the process of extensional deformation, the increase in the number and length of faults was episodic, and every ＇episode＇ experienced three periods, strain-accumulation period, quick fault-increment period and strain-adjustment period. The more complex the shape of the boundary fault, the higher the strain increment each ＇episode＇ experienced. Different extensional modes resulted in different fault-increment patterns. The horizontal detachment extensional mode has the ＇linear＇ style of fault-increment pattern, while the extensional mode controlled by a listric fault has the ＇stepwise＇ style of fault-increment pattern, and the extensional mode controlled by a ramp-flat boundary fault has the ＇stepwise-linear＇ style of fault-increment pattern. These fault-increment patterns given above could provide a theoretical method of fault interpretation and fracture prediction in extensional basins.展开更多
A sandbox experiment model was designed to simulate how differences in rock strength and gravity between two blocks can influence the formation characteristics of thrusts. In the experiment the compression was from on...A sandbox experiment model was designed to simulate how differences in rock strength and gravity between two blocks can influence the formation characteristics of thrusts. In the experiment the compression was from one direction with basement shortening and the initial surfaces of the model were oblique. The results show that if the initial surface was horizontal or the slope angle was smaller than 7°, the compression induced two groups of thrusts with opposite dip orientations. If the slope angle of the initial surface was greater than 7°, the compression induced only one group of thrusts with a dip orientation contrary to the original compression direction. This result is similar to the actual section of a collision zone between two continental blocks. By applying stress analysis, rock strength is shown to be an important factor in deformation. As other boundary conditions are changeless, it is the change of gravitational potential energy that leads to different deformation styles.展开更多
Ransomware is a type of malicious software that blocks access to a computer by encrypting user’s files until a ransom is paid to the attacker.There have been several reported high-profile ransomware attacks including...Ransomware is a type of malicious software that blocks access to a computer by encrypting user’s files until a ransom is paid to the attacker.There have been several reported high-profile ransomware attacks including WannaCry,Petya,and Bad Rabbit resulting in losses of over a billion dollars to various individuals and businesses in the world.The analysis of ransomware is often carried out via sandbox environments;however,the initial setup and configuration of such environments is a challenging task.Also,it is difficult for an ordinary computer user to correctly interpret the complex results presented in the reports generated by such environments and analysis tools.In this research work,we aim to develop a user-friendly model to understand the taxonomy and analysis of ransomware attacks.Also,we aim to present the results of analysis in the form of summarized reports that can easily be understood by an ordinary computer user.Our model is built on top of the well-known Cuckoo sandbox environment for identification of the ransomware as well as generation of the summarized reports.In addition,for evaluating the usability and accessibility of our proposed model,we conduct a comprehensive user survey consisting of participants from various fields,e.g.,professional developers from software houses,people from academia(professors,students).Our evaluation results demonstrate a positive feedback of approximately 92%on the usability of our proposed model.展开更多
This paper explores the expansion from fintech-based regulatory sandboxes to those that include medical artificial intelligence(AI)by examining their potential to foster innovation and accelerate market access while e...This paper explores the expansion from fintech-based regulatory sandboxes to those that include medical artificial intelligence(AI)by examining their potential to foster innovation and accelerate market access while ensuring safety and compliance,especially considering how they provide a flexible framework for medical AI companies to develop and test new technologies.This work also highlights the key risks involved,including data privacy,ethical concerns,real-world validation,and regulatory consistency,and proposes strategies for mitigation.Using case studies from the United States,the United Kingdom,the European Union,Indonesia,Japan,and China,this paper illustrates how regulatory sandboxes can support AI-driven healthcare innovation.Overall,although regulatory sandboxes present several risks,they are valuable for advancing medical AI,granted that they balance innovation with robust oversight to ensure patient safety and long-term benefits.展开更多
Studying contaminant transport in the capillary fringe(CF),a crucial part of the vadose zone,offers insights into the mechanisms controlling pollution in soils and groundwater aquifers.This paper investigated contamin...Studying contaminant transport in the capillary fringe(CF),a crucial part of the vadose zone,offers insights into the mechanisms controlling pollution in soils and groundwater aquifers.This paper investigated contaminant transport in the CF by continuously injecting a conservative tracer(NaCl)and graphene oxide nanoparticle(GONP),an adsorptive contaminant,into a sandbox.After entering the CF from the unsaturated zone,both NaCl and GONP underwent lateral transport.The breakthrough curves(BTCs)for NaCl and GONP were derived from water samples collected at predetermined sampling holes.Subsequently,contaminant transport in the CF was modeled using a one-dimensionaletwo-dimensional(1D-2D)coupled hydrodynamic model.This model incorporated lateral dispersivity(aL=1.198 cm)and longitudinal dispersivity(aT=0.286 cm),calculated using a point-by-point method.The hydrodynamic dispersion coefficients obtained were then applied to the Brooks and Corey(BC)and the van Genuchten(VG)parametric models.The BC model more accurately simulated the NaCl migration compared to the VG model,leading to its application in simulating GONP transport in the CF.However,the simulated BTCs for GONP showed a lag behind the measured data,especially at high ionic strengths.This discrepancy was attributed to the variable adsorption partition coefficient of GONP under different ionic conditions.During the experiment,GONP adsorption onto the porous media's surface altered the capillary dynamics,notably increasing capillary rise height,decreasing seepage velocity,and reducing GONP dispersion.Therefore,it is necessary to consider the adsorption capacity of the contaminants in order to accurately assess their transport within the vadose zone.展开更多
In recent years,cyber threats have escalated across diverse sectors,with cybercrime syndicates increasingly exploiting system vulnerabilities.Traditional passive defense mechanisms have proven insufficient,particularl...In recent years,cyber threats have escalated across diverse sectors,with cybercrime syndicates increasingly exploiting system vulnerabilities.Traditional passive defense mechanisms have proven insufficient,particularly as Linux platforms—historically overlooked in favor of Windows—have emerged as frequent targets.According to Trend Micro,there has been a substantial increase in Linux-targeted malware,with ransomware attacks on Linux surpassing those on macOS.This alarming trend underscores the need for detection strategies specifically designed for Linux environments.To address this challenge,this study proposes a comprehensive malware detection framework tailored for Linux systems,integrating dynamic behavioral analysis with the semantic reasoning capabilities of large language models(LLMs).Malware samples are executed within sandbox environments to extract behavioral features such as system calls and command-line executions.These features are then systematically mapped to the MITRE ATT&CK framework,incorporating its defined data sources,data components,and Tactics,Techniques,and Procedures(TTPs).Two mapping constructs—Conceptual Definition Mapping and TTP Technical Keyword Mapping—are developed from official MITRE documentation.These resources are utilized to fine-tune an LLM,enabling it to semantically interpret complex behavioral patterns and infer associated attack techniques,including those employed by previously unknown malware variants.The resulting detection pipeline effectively bridges raw behavioral data with structured threat intelligence.Experimental evaluations confirm the efficacy of the proposed system,with the fine-tuned Gemma 2B model demonstrating significantly enhanced accuracy in associating behavioral features with ATT&CK-defined techniques.This study contributes a fully integrated Linux-specific detection framework,a novel approach for transforming unstructured behavioral data into actionable intelligence,improved interpretability of malicious behavior,and a scalable training process for future applications of LLMs in cybersecurity.展开更多
文摘The main challenge in AI governance today is striking a balance between controlling AI dangers and fostering AI innovation.Regulators in a number of nations have progressively extended the regulatory sandbox,which was first implemented in the banking sector,to AI governance in an effort to reduce the conflict between regulation and innovation.The AI regulatory sandbox is a new and feasible route for AI governance in China that not only helps to manage the risks of technology application but also prevents inhibiting AI innovation.It keeps inventors'trial-and-error tolerance space inside the regulatory purview while offering a controlled setting for the development and testing of novel AI that hasn't yet been put on the market.By providing full-cycle governance of AI with the principles of agility and inclusive prudence,the regulatory sandbox offers an alternative to the conventional top-down hard regulation,expost regulation,and tight regulation.However,the current system also has inherent limitations and practical obstacles that need to be overcome by a more rational and effective approach.To achieve its positive impact on AI governance,the AI regulatory sandbox system should build and improve the access and exit mechanism,the coordination mechanism between the sandbox and personal information protection,and the mechanisms of exemption,disclosure,and communication.
基金supported by China National Major Project of Oil and Gas (2011ZX05023-004-012, 2011ZX05006-006-02-01)China Natural Science Foundation (Grant No. 40772086)
文摘Fault formation and evolution in the presence of multiple pre-existing weaknesses has not been investigated extensively in rift basins. The fault systems of Weixinan Sag, Beibuwan Basin of China, which is fully covered with high-precision 3-D seismic data and is rich in oil-gas resources, have been successfully reproduced by sandbox modeling in this study with inclusion of multiple pre-existing weaknesses in the experimental model. The basic characteristics of fault formation and evolution revealed by sandbox modeling are as follows. 1） Weakness-reactivation faults and weakness-related faults are formed much earlier than the distant-weakness faults （faults far away from and with little or no relationship to the weakness）. 2） Weakness-reactivation faults and weakness-related faults develop mainly along or parallel to a pre-existing weakness, while distant-weakness faults develop nearly perpendicular to the extension direction. A complicated fault system can be formed in a fixed direction of extension with the existence of multiple pre-existing weaknesses, and the complicated fault system in the Weixinan Sag formed gradually in a nearly N-S direction with multiple pre-existing weaknesses. 3） The increase in the length and number of faults is closely tied to the nature of pre-existing weaknesses. The sandbox model may provide a new clue to detailed fault system research for oil and gas exploration in rift basins.
文摘Three series of sandbox modeling experiments were performed to study the fault-increment pattern in extensional basins. Experimental results showed that the tectonic action mode of boundaries and the shape of major boundary faults control the formation and evolution of faults in extensional basins. In the process of extensional deformation, the increase in the number and length of faults was episodic, and every ＇episode＇ experienced three periods, strain-accumulation period, quick fault-increment period and strain-adjustment period. The more complex the shape of the boundary fault, the higher the strain increment each ＇episode＇ experienced. Different extensional modes resulted in different fault-increment patterns. The horizontal detachment extensional mode has the ＇linear＇ style of fault-increment pattern, while the extensional mode controlled by a listric fault has the ＇stepwise＇ style of fault-increment pattern, and the extensional mode controlled by a ramp-flat boundary fault has the ＇stepwise-linear＇ style of fault-increment pattern. These fault-increment patterns given above could provide a theoretical method of fault interpretation and fracture prediction in extensional basins.
基金This paper is supported by the project IGCP411(3-3-02-24) .
文摘A sandbox experiment model was designed to simulate how differences in rock strength and gravity between two blocks can influence the formation characteristics of thrusts. In the experiment the compression was from one direction with basement shortening and the initial surfaces of the model were oblique. The results show that if the initial surface was horizontal or the slope angle was smaller than 7°, the compression induced two groups of thrusts with opposite dip orientations. If the slope angle of the initial surface was greater than 7°, the compression induced only one group of thrusts with a dip orientation contrary to the original compression direction. This result is similar to the actual section of a collision zone between two continental blocks. By applying stress analysis, rock strength is shown to be an important factor in deformation. As other boundary conditions are changeless, it is the change of gravitational potential energy that leads to different deformation styles.
基金support of Security Testing-Innovative Secured Systems Lab(ISSL)established at University of Engineering&Technology,Peshawar,Pakistan under the Higher Education Commission initiative of National Center for Cyber Security(Grant No.2(1078)/HEC/M&E/2018/707).
文摘Ransomware is a type of malicious software that blocks access to a computer by encrypting user’s files until a ransom is paid to the attacker.There have been several reported high-profile ransomware attacks including WannaCry,Petya,and Bad Rabbit resulting in losses of over a billion dollars to various individuals and businesses in the world.The analysis of ransomware is often carried out via sandbox environments;however,the initial setup and configuration of such environments is a challenging task.Also,it is difficult for an ordinary computer user to correctly interpret the complex results presented in the reports generated by such environments and analysis tools.In this research work,we aim to develop a user-friendly model to understand the taxonomy and analysis of ransomware attacks.Also,we aim to present the results of analysis in the form of summarized reports that can easily be understood by an ordinary computer user.Our model is built on top of the well-known Cuckoo sandbox environment for identification of the ransomware as well as generation of the summarized reports.In addition,for evaluating the usability and accessibility of our proposed model,we conduct a comprehensive user survey consisting of participants from various fields,e.g.,professional developers from software houses,people from academia(professors,students).Our evaluation results demonstrate a positive feedback of approximately 92%on the usability of our proposed model.
基金National Science and Technology Major Project(2020AAA0105000)Beijing Science and Technology Plan(Z241100007724003).
文摘This paper explores the expansion from fintech-based regulatory sandboxes to those that include medical artificial intelligence(AI)by examining their potential to foster innovation and accelerate market access while ensuring safety and compliance,especially considering how they provide a flexible framework for medical AI companies to develop and test new technologies.This work also highlights the key risks involved,including data privacy,ethical concerns,real-world validation,and regulatory consistency,and proposes strategies for mitigation.Using case studies from the United States,the United Kingdom,the European Union,Indonesia,Japan,and China,this paper illustrates how regulatory sandboxes can support AI-driven healthcare innovation.Overall,although regulatory sandboxes present several risks,they are valuable for advancing medical AI,granted that they balance innovation with robust oversight to ensure patient safety and long-term benefits.
基金funded by the Yangzhou Talent Program“LvYangJingFeng”(YZLYJFJH2022YXBS124).
文摘Studying contaminant transport in the capillary fringe(CF),a crucial part of the vadose zone,offers insights into the mechanisms controlling pollution in soils and groundwater aquifers.This paper investigated contaminant transport in the CF by continuously injecting a conservative tracer(NaCl)and graphene oxide nanoparticle(GONP),an adsorptive contaminant,into a sandbox.After entering the CF from the unsaturated zone,both NaCl and GONP underwent lateral transport.The breakthrough curves(BTCs)for NaCl and GONP were derived from water samples collected at predetermined sampling holes.Subsequently,contaminant transport in the CF was modeled using a one-dimensionaletwo-dimensional(1D-2D)coupled hydrodynamic model.This model incorporated lateral dispersivity(aL=1.198 cm)and longitudinal dispersivity(aT=0.286 cm),calculated using a point-by-point method.The hydrodynamic dispersion coefficients obtained were then applied to the Brooks and Corey(BC)and the van Genuchten(VG)parametric models.The BC model more accurately simulated the NaCl migration compared to the VG model,leading to its application in simulating GONP transport in the CF.However,the simulated BTCs for GONP showed a lag behind the measured data,especially at high ionic strengths.This discrepancy was attributed to the variable adsorption partition coefficient of GONP under different ionic conditions.During the experiment,GONP adsorption onto the porous media's surface altered the capillary dynamics,notably increasing capillary rise height,decreasing seepage velocity,and reducing GONP dispersion.Therefore,it is necessary to consider the adsorption capacity of the contaminants in order to accurately assess their transport within the vadose zone.
基金supported by the National Science and Technology Council under grant number 113-2221-E-027-126-MY3.
文摘In recent years,cyber threats have escalated across diverse sectors,with cybercrime syndicates increasingly exploiting system vulnerabilities.Traditional passive defense mechanisms have proven insufficient,particularly as Linux platforms—historically overlooked in favor of Windows—have emerged as frequent targets.According to Trend Micro,there has been a substantial increase in Linux-targeted malware,with ransomware attacks on Linux surpassing those on macOS.This alarming trend underscores the need for detection strategies specifically designed for Linux environments.To address this challenge,this study proposes a comprehensive malware detection framework tailored for Linux systems,integrating dynamic behavioral analysis with the semantic reasoning capabilities of large language models(LLMs).Malware samples are executed within sandbox environments to extract behavioral features such as system calls and command-line executions.These features are then systematically mapped to the MITRE ATT&CK framework,incorporating its defined data sources,data components,and Tactics,Techniques,and Procedures(TTPs).Two mapping constructs—Conceptual Definition Mapping and TTP Technical Keyword Mapping—are developed from official MITRE documentation.These resources are utilized to fine-tune an LLM,enabling it to semantically interpret complex behavioral patterns and infer associated attack techniques,including those employed by previously unknown malware variants.The resulting detection pipeline effectively bridges raw behavioral data with structured threat intelligence.Experimental evaluations confirm the efficacy of the proposed system,with the fine-tuned Gemma 2B model demonstrating significantly enhanced accuracy in associating behavioral features with ATT&CK-defined techniques.This study contributes a fully integrated Linux-specific detection framework,a novel approach for transforming unstructured behavioral data into actionable intelligence,improved interpretability of malicious behavior,and a scalable training process for future applications of LLMs in cybersecurity.
