Due to the postponement of the conference, we will be updating the special sessions to the availability of our faculty at the new scheduled dates. Please visit this page again soon.

SPECIAL SESSIONS AND WORKSHOP

Special Session :
Advances in Visualization of Geotechnical Processes through Computer Vision and Deep Learning Technologies
Session Chair : 
Ningjun JIANG, University of Hawaii at Manoa 
jiangn@hawaii.edu

As a traditional civil engineering subject, geotechnical engineering deals with geo-structures built on, in or with soils and rocks. The intrinsic uncertainty, complexity and invisibility of such materials have made the recognition and understanding of geotechnical processes challenging. Recently, with the revolutionary progresses in data collection, cyberinfrastructure, and algorithms, more sophisticated methods such as computer vision and deep learning technologies have been applied to digitize, visualize, and analyze geotechnical problems. This Special Session will cover, but not limited to, the following topics:

  • Photogrammetry and Lidar for 3D reconstruction of geo-structures and features

  • 3D geotechnical structure detection from point clouds

  • Virtual reality of 3D geotechnical and geological models

  • Deep-learning based image processing for geotechnical and geological features

    recognition, segmentation, and quantification

  • Convolutional neural network (CNN) for sensor data extraction, identification, and

    processing

Special Session :
Machine Learning and Computational Methods applied to Ground Improvement​
Session Chair : 
Shuilong SHEN, Shantou University 
shensl@stu.edu.cn
Guiseppe MODONI, University of Cassino and Southern Lazio    
modoni@unicas.it
Session Secretaries : 
Khalid ELBAZ, Shantou University 
khalid@stu.edu.cn
Pierre Guy Atangana Njock 
atangana@stju.edu.cn  

The ground-improvement technology has greatly progressed in the last decades, providing unparalleled solutions to increase strength and stiffness of soft ground, modify water conductivity, prevent liquefaction etc., stimulating in this way the fantasy of designers for innovative application. Conversely, design too often relies on simplistic analyses, believed to be conservative but lacking of robust scientific background. Such a mismatch mostly stems from the inability to predict the effects of technology like in typical industrial processes, being their relation with the executive operation affected by epistemic and aleatory uncertainties on subsoil composition, inaccurate technology. Machine learning tools have become very popular to infer relations among variables in complex processes, weighting each factor straightly from observation, getting rid in this way of the subjective judgement of developers. Coupled with numerical computation, they offer powerful tools, alternative or complementary to classical analytical models, to improve reliability of design, forecasting the ground improved response from the geological conditions and construction operation and parameters.

The recent literature offers plenty examples of machine learning techniques referred to geotechnical application and, more specifically, to ground improvement. Aim of the proposed session is to gather all recent development of machine learning techniques applied to ground improvement, diffuse by contamination the different adopted methods among researchers and stimulate the coupling with computational methods to quantify their efficacy.

 

We set the following topics but not limited in these topics.

  • Data mining in ground improvement;

  • Application of machine learning or deep learning techniques in the prediction of ground response.

  • Development of hybrid meta-heuristic algorithms in the improvement of ground response.

  • Performance prediction of ground surface settlement through time series data mining.

  • Mechanism of ground improvement methods.

  • Ground improvement works in soft ground and related factors.

  • Analysis and design of ground improvement in deep excavations.

  • Dynamic prediction of deep mixing column properties in soft soil.

  • Ground improvement in tunnelling.

Mini-Symposium :
Machine Learning in Geotechnics​
Session Chair : 
Wengang ZHANG, Chongqing University
zhangwg@cqu.edu.cn
Carlos ACOSTA, Land Transport Authority (LTA) Singapore
atsoca.c@outlook.sg
Zhongqiang LIU, Norwegian Geotechnical Institute 
zhongqiang.liu@ngi.no

This Mini-symposium will discuss the challenges, opportunities, and trends related to the adoption of Machine Learning & Big Data in Geotechnics research and industrial workflows. Topics relevant to the Mini-Symposium include, but are not limited to, analytical and numerical developments for:

  • Machine learning methods for inverse problem analysis

  • Machine learning methods for simulation of stochastic processes

  • Machine learning methods for the random field modeling of heterogeneous geomaterials

  • Machine learning for visualization of uncertainty

  • Big data and cloud computing analytics for the management of large scale geo-

    applications

  • Bayesian methods to improve Geo- engineering and scientific decision-making

  • Risk assessment and management via machine learning and big data

Special Session :
Applications of Soft Computing in Geotechnical Case Studies ​
Session Chair : 
Wengang ZHANG, Chongqing University 
zhangwg@cqu.edu.cn
Anthony TECK Chee Goh, Nanyang Technological University
ctcgoh@ntu.edu.sg
Lin WANG, Chongqing University 
sdxywanglin@cqu.edu.cn
Runhong ZHANG, Chongqing University
zhangrh@cqu.edu.cn

This special session will discuss the challenges, opportunities, and trends related to the adoption of soft computing techniques in solving the geotechnical problems. Topics relevant to the Special Session include, but are not limited to, analytical and numerical developments for:

  • Soft computing-based decision-making framework for geotechnical disaster early warning

  • Geotechnical risk assessment or landslide susceptibility prediction using deep learning

  • Geotechnical applications of soft techniques in conjunction with other artificial intelligence algorithms

  • Soft computing method for time-series geotechnical problems

  • Machine learning of TBM data and its applications

Special Session :
Data Analytics in Geotechnical and Geological Engineering ​
Session Chair : 
Yu WANG, City University of Hong Kong 
yuwang@cityu.edu.hk
Wengang ZHANG, Chongqing University
zhangwg@cqu.edu.cn
Xiaohui QI, Nanyang Technological University, Singapore 
qixiaohui@ntu.edu.sg
Jianye CHING, National Taiwan University
jyching@gmail.com

With the fast development of measuring equipment and digitalization technologies, recent years have seen a rapid growth of data in geotechnical and geological engineering, such as in-situ and laboratory testing soil properties, pile load test data, and geophysical data. Yet, site-specific data that can be directly used for designs are still largely scarce. This deficiency of data poses great challenges to the design, construction works of geotechnical structures, and the selection of future site investigation schemes due to the large variabilities of the geotechnical and geological properties. For example, the spatial variabilities of geotechnical and geological properties are difficult to characterize with limited data. The location of the interface of geological formations or rock head in the area between boreholes normally has a huge uncertainty. The design scheme of geotechnical structures derived from limited data is not likely to be robust to measurement errors and variability of soil properties. The practical questions arising from these challenges include how to make use of a generic or global database to infer the soil property in a specific site, how to reliably estimate the geological and geotechnical property in unsampled locations and how to obtain an unbiased evaluation of the safety of geotechnical structures with limited data and, in turn, a reasonable design scheme of geotechnical structures. Previous efforts to address these issues primarily focused on deterministic methods. These deterministic methods do not well consider the inherent variability of geotechnical and geological properties and the error in spatial variability characterizations (termed as statistical uncertainty). As such, the attention has been gradually shifted from deterministic to probabilistic methods. In addition, there is a growing interest in the application of data-driven methods to geotechnical and geological engineering.

 

The aim of this Special Session is to incorporate the recent developments in data analytic methods or data-driven methods in geotechnical and geological engineering. The special issue will cover, but not limited to the following topics:

  •  Back analysis of geotechnical structures

  •  Big data and database processing

  •  Data-driven site investigation and interpolation methods

  •  Geostatistics and probabilistic site characterization

  •  Quantification of geological uncertainties

  •  Reliability-based design considering statistical uncertainty

  •  Digital geology and geological modelling

Workshop : 
Practical Application of Machine Learning in Geotechnics ​
Session Chair : 
Georg H. ERHARTER, Graz University of Technology
Thomas MARCHER, Graz University of Technology
Franz TSCHCHNIGG, Graz University of Technology
The last years have seen a rising interest in the application of Machine Learning for geotechnical problems. However, many applications are only of academic or theoretic nature and there is a lack of practical applications that may be used as a benchmark or as a working basis for future developments. Therefore, a benchmark test / data science competition will be organized before the workshop. The goal of the Workshop is then to share and discuss experiences that were gained during the competition