As the geoscience field evolves, artificial intelligence (AI) and machine learning (ML) techniques are becoming an essential pillar of innovation. This course is specifically designed for geoscientists interested in harnessing the power of AI and ML to tackle practical challenges in their work.
Over two days, participants will gain a foundational understanding of AI and ML principles, coupled with hands-on experience to apply these techniques to real-world geoscientific problems. Python programming language will be intensively used.
The course will kick off with an introduction to the fundamental principles of AI and machine learning. Participants will learn about the workflow involved in implementing AI/ML solutions, from data collection and preparation to model training and evaluation. This foundational knowledge is crucial for understanding how AI/ML can be effectively applied in geoscience contexts.
We will delve into how AI and ML techniques can specifically address “geos” issues. Whether it’s predicting well production profiles or extracting geological information from text, the applications are vast and varied. By understanding the unique challenges faced by geoscientists, participants will be better equipped to implement AI/ML solutions that are not only innovative but also practical and applicable to their specific fields.
A significant portion of the course will be dedicated to hands-on learning in the AI/ML playground. Participants will engage in practical exercises covering essential topics such as:
Data Standardization: understanding the importance of uniform data formats and structures for effective analysis;
Data Balancing: techniques to handle imbalanced datasets, ensuring more reliable model performance;
Dataset Size: strategies for determining the optimal size of datasets for training models;
Overfitting: recognizing and addressing overfitting issues to enhance model generalization.
We will also cover both regular machine learning techniques and deep learning methodologies. Through practical exercises focused on well data, participants will gain experience in applying these techniques to real datasets.
To solidify learnings, participants will explore two specific use cases from the following list:
Rock/facies image recognition using a pre-trained Convolutional Neural Network;
Well production profile prediction;
Machine learning on well data for facies prediction;
Geological information extraction from text using Natural Language Processing (NLP) techniques.
Following the exploration of these use cases and if the course format allows (in presence), participants will engage in a team contest focused on one of the selected items. This collaborative exercise will involve data collection, data preparation, and the practical application of AI/ML techniques to train and evaluate models. Teams will present their findings, showcasing their understanding of the course material and the innovative solutions they have developed. This hands-on approach reinforces theoretical concepts and fosters teamwork and collaborative problem-solving.
By the end of this course, participants will have acquired valuable insights into AI and machine learning techniques tailored for the geoscience industry. They will be equipped with the skills needed to implement these powerful tools in their work, driving efficiency and innovation in their endeavors.
Participants’ Profile
This course is designed for geoscientists, engineers, and data enthusiasts who are looking to explore the potential of artificial intelligence (AI) and machine learning (ML) in the geoscience industry. It is ideal for those who have little to no prior experience with AI/ML but are eager to learn how these technologies can help them analyze complex datasets, automate workflows, and solve real-world problems. Participants should have a curiosity for data analysis and a willingness to engage in hands-on exercises that will enable them to apply AI/ML techniques directly to their field of work
Prerequisites
Participants should have a basic understanding of Python programming, as the course involves coding exercises and practical applications of AI/ML techniques. While advanced Python skills are not necessarily required, familiarity with data manipulation and visualization using Python libraries such as Pandas, GeoPandas, Matplotlib, Plotly,and Seaborn will definitely help participants get the most out of the course.
About the Instructor
Claude Cavelius holds a Master’s degree in Numerical Geology from the École Normale Supérieure de Géologie (Nancy, France), earned in 2007. A geologist by training, Claude has always been passionate about software development, technology and innovation. He began his career at Chevron, where he spent 9 years as a software engineer and research geologist. During this time, he specialized in geostatistics and structural geology, contributing to the development of advanced geological models and tools to support exploration and production activities. Claude’s dual expertise in geology and programming allowed him to bridge the gap between complex geoscientific challenges and efficient software solutions.
In 2016, Claude joined Belmont Technology as the product manager, where he focused on delivering advanced, cloud-based AI solutions tailored for the oil and gas industry. His role involved designing and implementing AI-driven tools that enabled more efficient data analysis and decision-making, helping clients optimize their operations through innovative technology.
Today, Claude serves as the CEO/CTO of DeepLime. DeepLime operates at the the crossroads of geology, IT, and data science, empowering businesses by unlocking the full potential of their geological data. Claude leads the software development team, which focuses on creating cutting-edge tools and solutions that transform the way geoscientists work.
