Through a series of studies on arithmetic coding and arithmetic encryption, a novel image joint compression- encryption algorithm based on adaptive arithmetic coding is proposed. The contexts produced in the process o...Through a series of studies on arithmetic coding and arithmetic encryption, a novel image joint compression- encryption algorithm based on adaptive arithmetic coding is proposed. The contexts produced in the process of image compression are modified by keys in order to achieve image joint compression encryption. Combined with the bit-plane coding technique, the discrete wavelet transform coefficients in different resolutions can be encrypted respectively with different keys, so that the resolution selective encryption is realized to meet different application needs. Zero-tree coding is improved, and adaptive arithmetic coding is introduced. Then, the proposed joint compression-encryption algorithm is simulated. The simulation results show that as long as the parameters are selected appropriately, the compression efficiency of proposed image joint compression-encryption algorithm is basically identical to that of the original image compression algorithm, and the security of the proposed algorithm is better than the joint encryption algorithm based on interval splitting.展开更多
This paper proposes an efficient lossless image compression scheme for still images based on an adaptive arithmetic coding compression algorithm. The algorithm increases the image coding compression rate and ensures t...This paper proposes an efficient lossless image compression scheme for still images based on an adaptive arithmetic coding compression algorithm. The algorithm increases the image coding compression rate and ensures the quality of the decoded image combined with the adaptive probability model and predictive coding. The use of adaptive models for each encoded image block dynamically estimates the probability of the relevant image block. The decoded image block can accurately recover the encoded image according to the code book information. We adopt an adaptive arithmetic coding algorithm for image compression that greatly improves the image compression rate. The results show that it is an effective compression technology.展开更多
This paper presents a new method of lossless image compression. An image is characterized by homogeneous parts. The bit planes, which are of high weight are characterized by sequences of 0 and 1 are successive encoded...This paper presents a new method of lossless image compression. An image is characterized by homogeneous parts. The bit planes, which are of high weight are characterized by sequences of 0 and 1 are successive encoded with RLE, whereas the other bit planes are encoded by the arithmetic coding (AC) (static or adaptive model). By combining an AC (adaptive or static) with the RLE, a high degree of adaptation and compression efficiency is achieved. The proposed method is compared to both static and adaptive model. Experimental results, based on a set of 12 gray-level images, demonstrate that the proposed scheme gives mean compression ratio that are higher those compared to the conventional arithmetic encoders.展开更多
Due to the development of CT (Computed Tomography), MRI (Magnetic Resonance Imaging), PET (Positron Emission Tomography), EBCT (Electron Beam Computed Tomography), SMRI (Stereotactic Magnetic Resonance Imaging), etc. ...Due to the development of CT (Computed Tomography), MRI (Magnetic Resonance Imaging), PET (Positron Emission Tomography), EBCT (Electron Beam Computed Tomography), SMRI (Stereotactic Magnetic Resonance Imaging), etc. has enhanced the distinguishing rate and scanning rate of the imaging equipments. The diagnosis and the process of getting useful information from the image are got by processing the medical images using the wavelet technique. Wavelet transform has increased the compression rate. Increasing the compression performance by minimizing the amount of image data in the medical images is a critical task. Crucial medical information like diagnosing diseases and their treatments is obtained by modern radiology techniques. Medical Imaging (MI) process is used to acquire that information. For lossy and lossless image compression, several techniques were developed. Image edges have limitations in capturing them if we make use of the extension of 1-D wavelet transform. This is because wavelet transform cannot effectively transform straight line discontinuities, as well geographic lines in natural images cannot be reconstructed in a proper manner if 1-D transform is used. Differently oriented image textures are coded well using Curvelet Transform. The Curvelet Transform is suitable for compressing medical images, which has more curvy portions. This paper describes a method for compression of various medical images using Fast Discrete Curvelet Transform based on wrapping technique. After transformation, the coefficients are quantized using vector quantization and coded using arithmetic encoding technique. The proposed method is tested on various medical images and the result demonstrates significant improvement in performance parameters like Peak Signal to Noise Ratio (PSNR) and Compression Ratio (CR).展开更多
The rapid advancement of three-dimensional printed concrete(3DPC)requires intelligent and interpretable frameworks to optimize mixture design for strength,printability,and sustainability.While machine learning(ML)mode...The rapid advancement of three-dimensional printed concrete(3DPC)requires intelligent and interpretable frameworks to optimize mixture design for strength,printability,and sustainability.While machine learning(ML)models have improved predictive accuracy,their limited transparency has hindered their widespread adoption in materials engineering.To overcome this barrier,this study introduces a Random Forests ensemble learning model integrated with SHapley Additive exPlanations(SHAP)and Partial Dependence Plots(PDPs)to model and explain the compressive strength behavior of 3DPC mixtures.Unlike conventional“black-box”models,SHAP quantifies each variable’s contribution to predictions based on cooperative game theory,which enables causal interpretability,whereas PDP visualizes nonlinear and interactive effects between features that offer practical mix design insights.A systematically optimized random forest model achieved strong generalization(R2=0.978 for training,0.834 for validation,and 0.868 for testing).The analysis identified curing age,Portland cement,silica fume,and the water-tobinder ratio as dominant predictors,with curing age exerting the highest positive influence on strength development.The integrated SHAP-PDP framework revealed synergistic interactions among binder constituents and curing parameters,which established transparent,data-driven guidelines for performance optimization.Theoretically,the study advances explainable artificial intelligence in cementitious material science by linking microstructural mechanisms to model-based reasoning,thereby enhancing both the interpretability and applicability of ML-driven mix design for next-generation 3DPC systems.展开更多
Accurately modeling real network dynamics is a grand challenge in network science.The network dynamics arise from node interactions,which are shaped by network topology.Real networks tend to exhibit compact or highly ...Accurately modeling real network dynamics is a grand challenge in network science.The network dynamics arise from node interactions,which are shaped by network topology.Real networks tend to exhibit compact or highly optimized topologies.But the key problems arise:how to compress a network to best enhance its compactness,and what the compression limit of the network reflects?We abstract the topological compression of complex networks as a dynamic process of making them more compact and propose the local compression modulus that plays a key role in effective compression evolution of networks.Subsequently,we identify topological compressibility-a general property of complex networks that characterizes the extent to which a network can be compressed-and provide its approximate quantification.We anticipate that our findings and established theory will provide valuable insights into both dynamics and various applications of complex networks.展开更多
The high temperature deformation behaviors of α＋β type titanium alloy TC11 （Ti-6.5Al-3.5Mo-1.5Zr-0.3Si） with coarse lamellar starting microstructure were investigated based on the hot compression tests in the tem...The high temperature deformation behaviors of α＋β type titanium alloy TC11 （Ti-6.5Al-3.5Mo-1.5Zr-0.3Si） with coarse lamellar starting microstructure were investigated based on the hot compression tests in the temperature range of 950-1100 ℃ and the strain rate range of 0.001-10 s-1. The processing maps at different strains were then constructed based on the dynamic materials model, and the hot compression process parameters and deformation mechanism were optimized and analyzed, respectively. The results show that the processing maps exhibit two domains with a high efficiency of power dissipation and a flow instability domain with a less efficiency of power dissipation. The types of domains were characterized by convergence and divergence of the efficiency of power dissipation, respectively. The convergent domain in a＋fl phase field is at the temperature of 950-990 ℃ and the strain rate of 0.001-0.01 s^-1, which correspond to a better hot compression process window of α＋β phase field. The peak of efficiency of power dissipation in α＋β phase field is at 950 ℃ and 0.001 s 1, which correspond to the best hot compression process parameters of α＋β phase field. The convergent domain in β phase field is at the temperature of 1020-1080 ℃ and the strain rate of 0.001-0.1 s^-l, which correspond to a better hot compression process window of β phase field. The peak of efficiency of power dissipation in ℃ phase field occurs at 1050 ℃ over the strain rates from 0.001 s^-1 to 0.01 s^-1, which correspond to the best hot compression process parameters of ,8 phase field. The divergence domain occurs at the strain rates above 0.5 s^-1 and in all the tested temperature range, which correspond to flow instability that is manifested as flow localization and indicated by the flow softening phenomenon in stress-- strain curves. The deformation mechanisms of the optimized hot compression process windows in a＋β and β phase fields are identified to be spheroidizing and dynamic recrystallizing controlled by self-diffusion mechanism, respectively. The microstructure observation of the deformed specimens in different domains matches very well with the optimized results.展开更多
Deep neural networks have achieved excellent classification results on several computer vision benchmarks.This has led to the popularity of machine learning as a service,where trained algorithms are hosted on the clou...Deep neural networks have achieved excellent classification results on several computer vision benchmarks.This has led to the popularity of machine learning as a service,where trained algorithms are hosted on the cloud and inference can be obtained on real-world data.In most applications,it is important to compress the vision data due to the enormous bandwidth and memory requirements.Video codecs exploit spatial and temporal correlations to achieve high compression ratios,but they are computationally expensive.This work computes the motion fields between consecutive frames to facilitate the efficient classification of videos.However,contrary to the normal practice of reconstructing the full-resolution frames through motion compensation,this work proposes to infer the class label from the block-based computed motion fields directly.Motion fields are a richer and more complex representation of motion vectors,where each motion vector carries the magnitude and direction information.This approach has two advantages:the cost of motion compensation and video decoding is avoided,and the dimensions of the input signal are highly reduced.This results in a shallower network for classification.The neural network can be trained using motion vectors in two ways:complex representations and magnitude-direction pairs.The proposed work trains a convolutional neural network on the direction and magnitude tensors of the motion fields.Our experimental results show 20×faster convergence during training,reduced overfitting,and accelerated inference on a hand gesture recognition dataset compared to full-resolution and downsampled frames.We validate the proposed methodology on the HGds dataset,achieving a testing accuracy of 99.21%,on the HMDB51 dataset,achieving 82.54%accuracy,and on the UCF101 dataset,achieving 97.13%accuracy,outperforming state-of-the-art methods in computational efficiency.展开更多
This study presents and verifies a hybrid methodology for reliable determination of parameters in structural rheological models(Zener,Burgers,and Maxwell)describing the viscoelastic behavior of polyurethane specimens ...This study presents and verifies a hybrid methodology for reliable determination of parameters in structural rheological models(Zener,Burgers,and Maxwell)describing the viscoelastic behavior of polyurethane specimens manufactured using extrusion-based 3D printing.Through comprehensive testing,including cyclic compression at strain rates ranging from 0.12 to 120 mm/min(0%-15%strain)and creep/relaxation experiments(10%-30%strain),the lumped parameters were independently determined using both analytical and numerical solutions of the models’differential equations,followed by cross-verification in additional experiments.Numerical solutions for creep and relaxation problems were obtained using finite element analysis,with the three-parameter Mooney-Rivlin model and Prony series employed to simulate elastic and viscous stress components,respectively.Energy dissipation per cycle was quantified during cyclic compression tests.The results demonstrate that all three models adequately describe material behavior within the 0%-15%strain range across various strain rates.Comparative analysis revealed the Burgers model’s superior performance in characterizing creep and stress relaxation at low strain levels.While Zener and Burgers model parameters from uniaxial compression showed limited applicability for energy dissipation calculations,the generalized Maxwell model effectively captured viscoelastic properties across different strain rates.Notably,parameters derived from creep tests provided a more universal assessment of dissipative properties due to optimization based on characteristic curve regions.Both parameter sets described polyurethane’s elastic-hysteretic behavior with approximately 20%error,proving significantly more accurate than the linear strain-time dependence hypothesis.Finite element analysis(FEA)complemented numerical modeling by demonstrating that while the generalized Maxwell model effectively describes initial rapid stress-strain changes,FEA provides superior characterization of steady-state processes.This computational approach yields more physically representative results compared to simplified analytical solutions,despite certain limitations in transient analysis.展开更多
The biological stabilization of soil using microbially induced carbonate precipitation(MICP)employs ureolytic bacteria to precipitate calcium carbonate(CaCO3),which binds soil particles,enhancing strength,stiffness,an...The biological stabilization of soil using microbially induced carbonate precipitation(MICP)employs ureolytic bacteria to precipitate calcium carbonate(CaCO3),which binds soil particles,enhancing strength,stiffness,and erosion resistance.The unconfinedcompressive strength(UCS),a key measure of soil strength,is critical in geotechnical engineering as it directly reflectsthe mechanical stability of treated soils.This study integrates explainable artificialintelligence(XAI)with geotechnical insights to model the UCS of MICP-treated sands.Using 517 experimental data points and a combination of various input variables—including median grain size(D50),coefficientof uniformity(Cu),void ratio(e),urea concentration(Mu),calcium concentration(Mc),optical density(OD)of bacterial solution,pH,and total injection volume(Vt)—fivemachine learning(ML)models,including eXtreme gradient boosting(XGBoost),Light gradient boosting machine(LightGBM),random forest(RF),gene expression programming(GEP),and multivariate adaptive regression splines(MARS),were developed and optimized.The ensemble models(XGBoost,LightGBM,and RF)were optimized using the Chernobyl disaster optimizer(CDO),a recently developed metaheuristic algorithm.Of these,LightGBM-CDO achieved the highest accuracy for UCS prediction.XAI techniques like feature importance analysis(FIA),SHapley additive exPlanations(SHAP),and partial dependence plots(PDPs)were also used to investigate the complex non-linear relationships between the input and output variables.The results obtained have demonstrated that the XAI-driven models can enhance the predictive accuracy and interpretability of MICP processes,offering a sustainable pathway for optimizing geotechnical applications.展开更多
基金supported by the Natural Science Foundation of Hainan Province, China (Grant No. 613155)
文摘Through a series of studies on arithmetic coding and arithmetic encryption, a novel image joint compression- encryption algorithm based on adaptive arithmetic coding is proposed. The contexts produced in the process of image compression are modified by keys in order to achieve image joint compression encryption. Combined with the bit-plane coding technique, the discrete wavelet transform coefficients in different resolutions can be encrypted respectively with different keys, so that the resolution selective encryption is realized to meet different application needs. Zero-tree coding is improved, and adaptive arithmetic coding is introduced. Then, the proposed joint compression-encryption algorithm is simulated. The simulation results show that as long as the parameters are selected appropriately, the compression efficiency of proposed image joint compression-encryption algorithm is basically identical to that of the original image compression algorithm, and the security of the proposed algorithm is better than the joint encryption algorithm based on interval splitting.
基金supported by the National Natural Science Foundation of China (Grant Nos. 60573172 and 60973152)the Superior University Doctor Subject Special Scientific Research Foundation of China (Grant No. 20070141014)the Natural Science Foundation of Liaoning Province of China (Grant No. 20082165)
文摘This paper proposes an efficient lossless image compression scheme for still images based on an adaptive arithmetic coding compression algorithm. The algorithm increases the image coding compression rate and ensures the quality of the decoded image combined with the adaptive probability model and predictive coding. The use of adaptive models for each encoded image block dynamically estimates the probability of the relevant image block. The decoded image block can accurately recover the encoded image according to the code book information. We adopt an adaptive arithmetic coding algorithm for image compression that greatly improves the image compression rate. The results show that it is an effective compression technology.
文摘This paper presents a new method of lossless image compression. An image is characterized by homogeneous parts. The bit planes, which are of high weight are characterized by sequences of 0 and 1 are successive encoded with RLE, whereas the other bit planes are encoded by the arithmetic coding (AC) (static or adaptive model). By combining an AC (adaptive or static) with the RLE, a high degree of adaptation and compression efficiency is achieved. The proposed method is compared to both static and adaptive model. Experimental results, based on a set of 12 gray-level images, demonstrate that the proposed scheme gives mean compression ratio that are higher those compared to the conventional arithmetic encoders.
文摘Due to the development of CT (Computed Tomography), MRI (Magnetic Resonance Imaging), PET (Positron Emission Tomography), EBCT (Electron Beam Computed Tomography), SMRI (Stereotactic Magnetic Resonance Imaging), etc. has enhanced the distinguishing rate and scanning rate of the imaging equipments. The diagnosis and the process of getting useful information from the image are got by processing the medical images using the wavelet technique. Wavelet transform has increased the compression rate. Increasing the compression performance by minimizing the amount of image data in the medical images is a critical task. Crucial medical information like diagnosing diseases and their treatments is obtained by modern radiology techniques. Medical Imaging (MI) process is used to acquire that information. For lossy and lossless image compression, several techniques were developed. Image edges have limitations in capturing them if we make use of the extension of 1-D wavelet transform. This is because wavelet transform cannot effectively transform straight line discontinuities, as well geographic lines in natural images cannot be reconstructed in a proper manner if 1-D transform is used. Differently oriented image textures are coded well using Curvelet Transform. The Curvelet Transform is suitable for compressing medical images, which has more curvy portions. This paper describes a method for compression of various medical images using Fast Discrete Curvelet Transform based on wrapping technique. After transformation, the coefficients are quantized using vector quantization and coded using arithmetic encoding technique. The proposed method is tested on various medical images and the result demonstrates significant improvement in performance parameters like Peak Signal to Noise Ratio (PSNR) and Compression Ratio (CR).
基金supported by the Ongoing Research Funding Program(Grant No.ORFFT-2025-025-4)at King Saud University,Riyadh,Saudi Arabia.The grant was awarded to Yassir M.Abbas。
文摘The rapid advancement of three-dimensional printed concrete(3DPC)requires intelligent and interpretable frameworks to optimize mixture design for strength,printability,and sustainability.While machine learning(ML)models have improved predictive accuracy,their limited transparency has hindered their widespread adoption in materials engineering.To overcome this barrier,this study introduces a Random Forests ensemble learning model integrated with SHapley Additive exPlanations(SHAP)and Partial Dependence Plots(PDPs)to model and explain the compressive strength behavior of 3DPC mixtures.Unlike conventional“black-box”models,SHAP quantifies each variable’s contribution to predictions based on cooperative game theory,which enables causal interpretability,whereas PDP visualizes nonlinear and interactive effects between features that offer practical mix design insights.A systematically optimized random forest model achieved strong generalization(R2=0.978 for training,0.834 for validation,and 0.868 for testing).The analysis identified curing age,Portland cement,silica fume,and the water-tobinder ratio as dominant predictors,with curing age exerting the highest positive influence on strength development.The integrated SHAP-PDP framework revealed synergistic interactions among binder constituents and curing parameters,which established transparent,data-driven guidelines for performance optimization.Theoretically,the study advances explainable artificial intelligence in cementitious material science by linking microstructural mechanisms to model-based reasoning,thereby enhancing both the interpretability and applicability of ML-driven mix design for next-generation 3DPC systems.
基金supported inpart by the National Natural Science Foundation of China(Grant No. 12371088)the Innovative Research Group Project of Natural Science Foundation of Hunan Provinceof China (Grant No. 2024JJ1008)in part by the Australian Research Council (ARC) through the Discovery Projects scheme (Grant No. DP220100580)。
文摘Accurately modeling real network dynamics is a grand challenge in network science.The network dynamics arise from node interactions,which are shaped by network topology.Real networks tend to exhibit compact or highly optimized topologies.But the key problems arise:how to compress a network to best enhance its compactness,and what the compression limit of the network reflects?We abstract the topological compression of complex networks as a dynamic process of making them more compact and propose the local compression modulus that plays a key role in effective compression evolution of networks.Subsequently,we identify topological compressibility-a general property of complex networks that characterizes the extent to which a network can be compressed-and provide its approximate quantification.We anticipate that our findings and established theory will provide valuable insights into both dynamics and various applications of complex networks.
基金Project (51005112) supported by the National Natural Science Foundation of ChinaProject (2010ZF56019) supported by the Aviation Science Foundation of China+1 种基金Project (GJJ11156) supported by the Education Commission of Jiangxi Province, ChinaProject(GF200901008) supported by the Open Fund of National Defense Key Disciplines Laboratory of Light Alloy Processing Science and Technology, China
文摘The high temperature deformation behaviors of α＋β type titanium alloy TC11 （Ti-6.5Al-3.5Mo-1.5Zr-0.3Si） with coarse lamellar starting microstructure were investigated based on the hot compression tests in the temperature range of 950-1100 ℃ and the strain rate range of 0.001-10 s-1. The processing maps at different strains were then constructed based on the dynamic materials model, and the hot compression process parameters and deformation mechanism were optimized and analyzed, respectively. The results show that the processing maps exhibit two domains with a high efficiency of power dissipation and a flow instability domain with a less efficiency of power dissipation. The types of domains were characterized by convergence and divergence of the efficiency of power dissipation, respectively. The convergent domain in a＋fl phase field is at the temperature of 950-990 ℃ and the strain rate of 0.001-0.01 s^-1, which correspond to a better hot compression process window of α＋β phase field. The peak of efficiency of power dissipation in α＋β phase field is at 950 ℃ and 0.001 s 1, which correspond to the best hot compression process parameters of α＋β phase field. The convergent domain in β phase field is at the temperature of 1020-1080 ℃ and the strain rate of 0.001-0.1 s^-l, which correspond to a better hot compression process window of β phase field. The peak of efficiency of power dissipation in ℃ phase field occurs at 1050 ℃ over the strain rates from 0.001 s^-1 to 0.01 s^-1, which correspond to the best hot compression process parameters of ,8 phase field. The divergence domain occurs at the strain rates above 0.5 s^-1 and in all the tested temperature range, which correspond to flow instability that is manifested as flow localization and indicated by the flow softening phenomenon in stress-- strain curves. The deformation mechanisms of the optimized hot compression process windows in a＋β and β phase fields are identified to be spheroidizing and dynamic recrystallizing controlled by self-diffusion mechanism, respectively. The microstructure observation of the deformed specimens in different domains matches very well with the optimized results.
基金Supported by Princess Nourah bint Abdulrahman University Researchers Supporting Project number(PNURSP2025R896).
文摘Deep neural networks have achieved excellent classification results on several computer vision benchmarks.This has led to the popularity of machine learning as a service,where trained algorithms are hosted on the cloud and inference can be obtained on real-world data.In most applications,it is important to compress the vision data due to the enormous bandwidth and memory requirements.Video codecs exploit spatial and temporal correlations to achieve high compression ratios,but they are computationally expensive.This work computes the motion fields between consecutive frames to facilitate the efficient classification of videos.However,contrary to the normal practice of reconstructing the full-resolution frames through motion compensation,this work proposes to infer the class label from the block-based computed motion fields directly.Motion fields are a richer and more complex representation of motion vectors,where each motion vector carries the magnitude and direction information.This approach has two advantages:the cost of motion compensation and video decoding is avoided,and the dimensions of the input signal are highly reduced.This results in a shallower network for classification.The neural network can be trained using motion vectors in two ways:complex representations and magnitude-direction pairs.The proposed work trains a convolutional neural network on the direction and magnitude tensors of the motion fields.Our experimental results show 20×faster convergence during training,reduced overfitting,and accelerated inference on a hand gesture recognition dataset compared to full-resolution and downsampled frames.We validate the proposed methodology on the HGds dataset,achieving a testing accuracy of 99.21%,on the HMDB51 dataset,achieving 82.54%accuracy,and on the UCF101 dataset,achieving 97.13%accuracy,outperforming state-of-the-art methods in computational efficiency.
文摘This study presents and verifies a hybrid methodology for reliable determination of parameters in structural rheological models(Zener,Burgers,and Maxwell)describing the viscoelastic behavior of polyurethane specimens manufactured using extrusion-based 3D printing.Through comprehensive testing,including cyclic compression at strain rates ranging from 0.12 to 120 mm/min(0%-15%strain)and creep/relaxation experiments(10%-30%strain),the lumped parameters were independently determined using both analytical and numerical solutions of the models’differential equations,followed by cross-verification in additional experiments.Numerical solutions for creep and relaxation problems were obtained using finite element analysis,with the three-parameter Mooney-Rivlin model and Prony series employed to simulate elastic and viscous stress components,respectively.Energy dissipation per cycle was quantified during cyclic compression tests.The results demonstrate that all three models adequately describe material behavior within the 0%-15%strain range across various strain rates.Comparative analysis revealed the Burgers model’s superior performance in characterizing creep and stress relaxation at low strain levels.While Zener and Burgers model parameters from uniaxial compression showed limited applicability for energy dissipation calculations,the generalized Maxwell model effectively captured viscoelastic properties across different strain rates.Notably,parameters derived from creep tests provided a more universal assessment of dissipative properties due to optimization based on characteristic curve regions.Both parameter sets described polyurethane’s elastic-hysteretic behavior with approximately 20%error,proving significantly more accurate than the linear strain-time dependence hypothesis.Finite element analysis(FEA)complemented numerical modeling by demonstrating that while the generalized Maxwell model effectively describes initial rapid stress-strain changes,FEA provides superior characterization of steady-state processes.This computational approach yields more physically representative results compared to simplified analytical solutions,despite certain limitations in transient analysis.
文摘The biological stabilization of soil using microbially induced carbonate precipitation(MICP)employs ureolytic bacteria to precipitate calcium carbonate(CaCO3),which binds soil particles,enhancing strength,stiffness,and erosion resistance.The unconfinedcompressive strength(UCS),a key measure of soil strength,is critical in geotechnical engineering as it directly reflectsthe mechanical stability of treated soils.This study integrates explainable artificialintelligence(XAI)with geotechnical insights to model the UCS of MICP-treated sands.Using 517 experimental data points and a combination of various input variables—including median grain size(D50),coefficientof uniformity(Cu),void ratio(e),urea concentration(Mu),calcium concentration(Mc),optical density(OD)of bacterial solution,pH,and total injection volume(Vt)—fivemachine learning(ML)models,including eXtreme gradient boosting(XGBoost),Light gradient boosting machine(LightGBM),random forest(RF),gene expression programming(GEP),and multivariate adaptive regression splines(MARS),were developed and optimized.The ensemble models(XGBoost,LightGBM,and RF)were optimized using the Chernobyl disaster optimizer(CDO),a recently developed metaheuristic algorithm.Of these,LightGBM-CDO achieved the highest accuracy for UCS prediction.XAI techniques like feature importance analysis(FIA),SHapley additive exPlanations(SHAP),and partial dependence plots(PDPs)were also used to investigate the complex non-linear relationships between the input and output variables.The results obtained have demonstrated that the XAI-driven models can enhance the predictive accuracy and interpretability of MICP processes,offering a sustainable pathway for optimizing geotechnical applications.
