Deep Learning(DL)techniques as a subfield of data science are getting overwhelming attention mainly because of their ability to understand the underlying pattern of data in making classifications.These techniques requ...Deep Learning(DL)techniques as a subfield of data science are getting overwhelming attention mainly because of their ability to understand the underlying pattern of data in making classifications.These techniques require a considerable amount of data to efficiently train the DL models.Generally,when the data size is larger,the DL models perform better.However,it is not possible to have a considerable amount of data in different domains such as healthcare.In healthcare,it is impossible to have a substantial amount of data to solve medical problems using Artificial Intelligence,mainly due to ethical issues and the privacy of patients.To solve this problem of small dataset,different techniques of data augmentation are used that can increase the size of the training set.However,these techniques only change the shape of the image and hence the classification model does not increase accuracy.Generative Adversarial Networks(GANs)are very powerful techniques to augment training data as new samples are created.This technique helps the classification models to increase their accuracy.In this paper,we have investigated augmentation techniques in healthcare image classification.The objective of this research paper is to develop a novel augmentation technique that can increase the size of the training set,to enable deep learning techniques to achieve higher accuracy.We have compared the performance of the image classifiers using the standard augmentation technique and GANs.Our results demonstrate that GANs increase the training data,and eventually,the classifier achieves an accuracy of 90%compared to standard data augmentation techniques,which achieve an accuracy of up to 70%.Other advanced CNN models are also tested and have demonstrated that more deep architectures can achieve more than 98%accuracy for making classification on Oral Squamous Cell Carcinoma.展开更多
The general use of aluminium as an indentation standard for the iteration of contact heights for the determination of ISO-14577 hardness and elastic modulus is challenged because of as yet not appreciated phase-change...The general use of aluminium as an indentation standard for the iteration of contact heights for the determination of ISO-14577 hardness and elastic modulus is challenged because of as yet not appreciated phase-changes in the physical force-depth standard curve that seemed to be secured by claims from 1992. The physical and mathematical analyses with closed formulas avoid the still world-wide standardized energy-law violation by not reserving 33.33% (h2 belief) (or 20% h3/2 physical law) of the loading force and thus energy for all not depth producing events but using 100% for the depth formation is a severe violation of the energy law. The not depth producing part of the indentation work cannot be done with zero energy! Both twinning and structural phase-transition onsets and normalized phase-transition energies are now calculated without iterations but with physically correct closed arithmetic equations. These are reported for Berkovich and cubecorner indentations, including their comparison on geometric grounds and an indentation standard without mechanical twinning is proposed. Characteristic data are reported. This is the first detection of the indentation twinning of aluminium at room temperature and the mechanical twinning of fused quartz is also new. Their disqualification as indentation standards is established. Also, the again found higher load phase-transitions disqualify aluminium and fused quartz as ISO-ASTM 14577 (International Standardization Organization and American Society for Testing and Materials) standards for the contact depth “hc” iterations. The incorrect and still world-wide used black-box values for H- and Er-values (the latter are still falsely called “Young’s moduli” even though they are not directional) and all mechanical properties that depend on them. They lack relation to bulk moduli from compression experiments. Experimentally obtained and so published force vs depth parabolas always follow the linear FN = kh3/2 + Fa equation, where Fa is the axis-cut before and after the phase-transition branches (never “h2” as falsely enforced and used for H, Er and giving incorrectly calculated parameters). The regression slopes k are the precise physical hardness values, which for the first time allow for precise calculation of the mechanical qualities by indentation in relation to the geometry of the indenter tip. Exactly 20% of the applied force and thus energy is not available for the indentation depth. Only these scientific k-values must be used for AI-advises at the expense of falsely iterated indentation hardness H-values. Any incorrect H-ISO-ASTM and also the iterated Er-ISO-ASTM modulus values of technical materials in artificial intelligence will be a disaster for the daily safety. The AI must be told that these are unscientific and must therefore be replaced by physical data. Iterated data (3 and 8 free parameters!) cannot be transformed into physical data. One has to start with real experimental loading curves and an absolute ZerodurR standard that must be calibrated with standard force and standard length to create absolute indentation results. .展开更多
In the wake of the research community gaining deep understanding about control-hijacking attacks,data-oriented attacks have emerged.Among data-oriented attacks,data structure manipulation attack(DSMA)is a major catego...In the wake of the research community gaining deep understanding about control-hijacking attacks,data-oriented attacks have emerged.Among data-oriented attacks,data structure manipulation attack(DSMA)is a major category.Pioneering research was conducted and shows that DSMA is able to circumvent the most effective defenses against control-hijacking attacks-DEP,ASLR and CFI.Up to this day,only two defense techniques have demonstrated their effectiveness:Data Flow Integrity(DFI)and Data Structure Layout Randomization(DSLR).However,DFI has high performance overhead,and dynamic DSLR has two main limitations.L-1:Randomizing a large set of data structures will significantly affect the performance.L-2:To be practical,only a fixed sub-set of data structures are randomized.In the case that the data structures targeted by an attack are not covered,dynamic DSLR is essentially noneffective.To address these two limitations,we propose a novel technique,feedback-control-based adaptive DSLR and build a system named SALADSPlus.SALADSPlus seeks to optimize the trade-off between security and cost through feedback control.Using a novel feedback-control-based adaptive algorithm extended from the Upper Confidence Bound(UCB)algorithm,the defender(controller)uses the feedbacks(cost-effectiveness)from previous randomization cycles to adaptively choose the set of data structures to randomize(the next action).Different from dynamic DSLR,the set of randomized data structures are adaptively changed based on the feedbacks.To obtain the feedbacks,SALADSPlus inserts canary in each data structure at the time of compilation.We have implemented SALADSPlus based on gcc-4.5.0.Experimental results show that the runtime overheads are 1.8%,3.7%,and 5.3% when the randomization cycles are selected as 10s,5s,and 1s respectively.展开更多
基金supported by Taif University Researchers Supporting Project No.(TURSP-2020/254),Taif University,Taif,Saudi Arabia.
文摘Deep Learning(DL)techniques as a subfield of data science are getting overwhelming attention mainly because of their ability to understand the underlying pattern of data in making classifications.These techniques require a considerable amount of data to efficiently train the DL models.Generally,when the data size is larger,the DL models perform better.However,it is not possible to have a considerable amount of data in different domains such as healthcare.In healthcare,it is impossible to have a substantial amount of data to solve medical problems using Artificial Intelligence,mainly due to ethical issues and the privacy of patients.To solve this problem of small dataset,different techniques of data augmentation are used that can increase the size of the training set.However,these techniques only change the shape of the image and hence the classification model does not increase accuracy.Generative Adversarial Networks(GANs)are very powerful techniques to augment training data as new samples are created.This technique helps the classification models to increase their accuracy.In this paper,we have investigated augmentation techniques in healthcare image classification.The objective of this research paper is to develop a novel augmentation technique that can increase the size of the training set,to enable deep learning techniques to achieve higher accuracy.We have compared the performance of the image classifiers using the standard augmentation technique and GANs.Our results demonstrate that GANs increase the training data,and eventually,the classifier achieves an accuracy of 90%compared to standard data augmentation techniques,which achieve an accuracy of up to 70%.Other advanced CNN models are also tested and have demonstrated that more deep architectures can achieve more than 98%accuracy for making classification on Oral Squamous Cell Carcinoma.
文摘The general use of aluminium as an indentation standard for the iteration of contact heights for the determination of ISO-14577 hardness and elastic modulus is challenged because of as yet not appreciated phase-changes in the physical force-depth standard curve that seemed to be secured by claims from 1992. The physical and mathematical analyses with closed formulas avoid the still world-wide standardized energy-law violation by not reserving 33.33% (h2 belief) (or 20% h3/2 physical law) of the loading force and thus energy for all not depth producing events but using 100% for the depth formation is a severe violation of the energy law. The not depth producing part of the indentation work cannot be done with zero energy! Both twinning and structural phase-transition onsets and normalized phase-transition energies are now calculated without iterations but with physically correct closed arithmetic equations. These are reported for Berkovich and cubecorner indentations, including their comparison on geometric grounds and an indentation standard without mechanical twinning is proposed. Characteristic data are reported. This is the first detection of the indentation twinning of aluminium at room temperature and the mechanical twinning of fused quartz is also new. Their disqualification as indentation standards is established. Also, the again found higher load phase-transitions disqualify aluminium and fused quartz as ISO-ASTM 14577 (International Standardization Organization and American Society for Testing and Materials) standards for the contact depth “hc” iterations. The incorrect and still world-wide used black-box values for H- and Er-values (the latter are still falsely called “Young’s moduli” even though they are not directional) and all mechanical properties that depend on them. They lack relation to bulk moduli from compression experiments. Experimentally obtained and so published force vs depth parabolas always follow the linear FN = kh3/2 + Fa equation, where Fa is the axis-cut before and after the phase-transition branches (never “h2” as falsely enforced and used for H, Er and giving incorrectly calculated parameters). The regression slopes k are the precise physical hardness values, which for the first time allow for precise calculation of the mechanical qualities by indentation in relation to the geometry of the indenter tip. Exactly 20% of the applied force and thus energy is not available for the indentation depth. Only these scientific k-values must be used for AI-advises at the expense of falsely iterated indentation hardness H-values. Any incorrect H-ISO-ASTM and also the iterated Er-ISO-ASTM modulus values of technical materials in artificial intelligence will be a disaster for the daily safety. The AI must be told that these are unscientific and must therefore be replaced by physical data. Iterated data (3 and 8 free parameters!) cannot be transformed into physical data. One has to start with real experimental loading curves and an absolute ZerodurR standard that must be calibrated with standard force and standard length to create absolute indentation results. .
基金supported by ARO W911NF-13-1-0421(MURI)NSF CNS-1422594NSF CNS-1505664.
文摘In the wake of the research community gaining deep understanding about control-hijacking attacks,data-oriented attacks have emerged.Among data-oriented attacks,data structure manipulation attack(DSMA)is a major category.Pioneering research was conducted and shows that DSMA is able to circumvent the most effective defenses against control-hijacking attacks-DEP,ASLR and CFI.Up to this day,only two defense techniques have demonstrated their effectiveness:Data Flow Integrity(DFI)and Data Structure Layout Randomization(DSLR).However,DFI has high performance overhead,and dynamic DSLR has two main limitations.L-1:Randomizing a large set of data structures will significantly affect the performance.L-2:To be practical,only a fixed sub-set of data structures are randomized.In the case that the data structures targeted by an attack are not covered,dynamic DSLR is essentially noneffective.To address these two limitations,we propose a novel technique,feedback-control-based adaptive DSLR and build a system named SALADSPlus.SALADSPlus seeks to optimize the trade-off between security and cost through feedback control.Using a novel feedback-control-based adaptive algorithm extended from the Upper Confidence Bound(UCB)algorithm,the defender(controller)uses the feedbacks(cost-effectiveness)from previous randomization cycles to adaptively choose the set of data structures to randomize(the next action).Different from dynamic DSLR,the set of randomized data structures are adaptively changed based on the feedbacks.To obtain the feedbacks,SALADSPlus inserts canary in each data structure at the time of compilation.We have implemented SALADSPlus based on gcc-4.5.0.Experimental results show that the runtime overheads are 1.8%,3.7%,and 5.3% when the randomization cycles are selected as 10s,5s,and 1s respectively.
