Suggested by the structure of the visual nervous system, a new algorithm is proposed for pattern recognition. This algorithm can be realized with a multilayered network consisting of neuron-like cells. The network, “neocognitron”, is self-organized by unsupervised learning, and acquires the ability to recognize stimulus patterns according to the differences in their shapes: Any patterns which we human beings judge to be alike are also judged to be of the same category by the neocognitron. The neocognitron recognizes stimulus patterns correctly without being affected by shifts in position or even by considerable distortions in shape of the stimulus patterns.

Neocognitron--A New Algorithm for Pattern Recognition Tolerant of Deformations and Shifts in Position

Accurate assessment of undrained shear strength (USS) for soft sensitive clays is a great concern in geotechnical engineering practice. This study applies novel data-driven extreme gradient boosting (XGBoost) and random forest (RF) ensemble learning methods for capturing the relationships between the USS and various basic soil parameters. Based on the soil data sets from TC304 database, a general approach is developed to predict the USS of soft clays using the two machine learning methods above, where five feature variables including the preconsolidation stress (PS), vertical effective stress (VES), liquid limit (LL), plastic limit (PL) and natural water content (W) are adopted. To reduce the dependence on the rule of thumb and inefficient brute-force search, the Bayesian optimization method is applied to determine the appropriate model hyper-parameters of both XGBoost and RF. The developed models are comprehensively compared with three comparison machine learning methods and two transformation models with respect to predictive accuracy and robustness under 5-fold cross-validation (CV). It is shown that XGBoost-based and RF-based methods outperform these approaches. Besides, the XGBoost-based model provides feature importance ranks, which makes it a promising tool in the prediction of geotechnical parameters and enhances the interpretability of model.

Prediction of undrained shear strength using extreme gradient boosting and random forest based on Bayesian optimization

With the advent of big data era, deep learning (DL) has become an essential research subject in the field of artificial intelligence (AI). DL algorithms are characterized with powerful feature learning and expression capabilities compared with the traditional machine learning (ML) methods, which attracts worldwide researchers from different fields to its increasingly wide applications. Furthermore, in the field of geochnical engineering, DL has been widely adopted in various research topics, a comprehensive review summarizing its application is desirable. Consequently, this study presented the state of practice of DL in geotechnical engineering, and depicted the statistical trend of the published papers. Four major algorithms, including feedforward neural (FNN), recurrent neural network (RNN), convolutional neural network (CNN) and generative adversarial network (GAN) along with their geotechnical applications were elaborated. In addition, a thorough summary containing pubilished literatures, the corresponding reference cases, the adopted DL algorithms as well as the related geotechnical topics was compiled. Furthermore, the challenges and perspectives of future development of DL in geotechnical engineering were presented and discussed.

Application of deep learning algorithms in geotechnical engineering: a short critical review

The effect of particle shape on the strength, dilatancy, and stress-dilatancy relationship was systematically investigated through a series of drained triaxial compression tests on sands mi...

https://ascelibrary.org/doi/pdf/10.1061/%28ASCE%29GT.1943-5606.0001994

Effect of Particle Shape on Stress-Dilatancy Responses of Medium-Dense Sands

Particle breakage has a great influence on stress-strain behaviors of rockfill materials. A new relative breakage was proposed for rockfill materials. A critical state line incorporating particle breakage was termed as a breakage critical state line (BCSL). It was found that the BCSL position of rockfill materials pertained to the relative breakage. An elastoplastic model considering state dependence and particle breakage was proposed for rockfill materials. The effect of particle breakage was incorporated into the proposed state parameter, dilatancy stress ratio, stress dilatancy relation, bounding stress ratio, and plastic modulus in the proposed model. The model considering particle breakage can well predict the stress-strain and particle-breakage behaviors of rockfill materials at various confining pressures.

Elastoplastic Constitutive Model for Rockfill Materials Considering Particle Breakage

The influences of particle breakage on the position of the critical state line (CSL) were systematically investigated in this paper through a series of large-scale triaxial compression tests on Tacheng rockfill material (TRM). It was found that the critical-state stress ratio of TRM (i.e., the gradient of the CSL in the p−q space) was approximately regarded as a constant. In the e−logp space, the CSL of TRM descended with a decrease in the initial void ratio, whereas the gradient of the CSL was constant. A procedure was established for evaluating the critical state point at a same particle breakage, which comprised the breakage critical state line (BCSL). An increase of the particle breakage led to not only a vertical translation but also a rotation on the BCSL of TRM in the e−logp space, which was similar to the observation of Dog’s Bay sand. Consequently, the initial gradation (or the corresponding initial void ratio) was the dominant factor that affected the position of the CSL of TRM in the e−...

Influence of Particle Breakage on Critical State Line of Rockfill Material

Assessment of soil liquefaction based on capacity energy concept and multivariate adaptive regression splines

The amount of particle breakage of Tacheng rockfill material (TRM) in the process of shearing was quantitatively measured by a relative breakage index. The fractal dimension of TRM at the final state of the tests was found to be linearly related to the void ratio and logarithm of the confining pressure. The influences of the relative breakage index on the friction angle, modulus, and deformation were investigated based on a series of large-scale triaxial compression tests on TRM. An increase in the relative breakage index leads to an increase in the volumetric strain and to decreases in the peak state friction angle, critical state friction angle, deviatoric strain, and critical state void ratio. The test results were used to propose simple formulations pertaining to the relative breakage index for the strength and deformation indexes (e.g., peak state friction angle, critical state friction angle, initial elastic modulus, secant modulus at 50% of the peak strength, volumetric strain, deviatoric s...

Strength and Deformation of Rockfill Material Based on Large-Scale Triaxial Compression Tests. II: Influence of Particle Breakage

A series of large-scale triaxial compression tests were conducted to investigate the strength and deformation behaviors of Tacheng rockfill material (TRM) in relation to the initial void ratio and initial confining pressure. The critical state friction angle of TRM was expressed as a linear function of the logarithm of the initial confining pressure. The excess peak state friction angle and excess characteristic state friction angle of TRM were formulated as linear equations of a revised relative dilatancy index to capture the influences of density and pressure on the peak state and characteristic state friction angles. The initial elastic modulus, tangent modulus, and secant modulus of TRM were dependent on the initial void ratio and initial confining pressure. In addition, a formulation incorporating density and pressure was proposed to simulate the initial elastic modulus of TRM. The volumetric and deviatoric strains of TRM at the critical state were also dependent on density and pressure. The ...

Yumin Chen

Papers

Application of deep learning algorithms in geotechnical engineering: a short critical review

Influence of Particle Breakage on Critical State Line of Rockfill Material

Assessment of soil liquefaction based on capacity energy concept and multivariate adaptive regression splines

Strength and Deformation of Rockfill Material Based on Large-Scale Triaxial Compression Tests. II: Influence of Particle Breakage

Strength and Deformation of Rockfill Material Based on Large-Scale Triaxial Compression Tests. I: Influences of Density and Pressure