24020 GeoAI-based Flood Susceptibility Mapping Using TAB Transformer Algorithms Optimized by Meta-Heuristic Techniques: A Climate Change Perspective

Project Title: 24020 GeoAI-based Flood Susceptibility Mapping Using TAB Transformer Algorithms Optimized by Meta-Heuristic Techniques: A Climate Change Perspective

Sponsorship: Spatial Data Lab for Geospatial AI Research and Environmental Hazard Mapping         

Project Description:

This project focuses on developing advanced flood susceptibility maps using GeoAI (Artificial Intelligence) methods, specifically leveraging Tab Transformer algorithms optimized by meta-heuristic techniques. As climate change intensifies flood risks globally, it becomes critical to employ cutting-edge technologies to predict flood-prone areas more accurately. By utilizing climate and geospatial data, this study integrates the strengths of Tab Transformer algorithms, known for handling mixed data types and complex relationships, with optimization techniques to enhance model performance. Meta-heuristic algorithms will be employed to optimize hyperparameters and model structures, ensuring robust flood risk predictions. The project will focus on obtaining and analyzing flood-related data from the U.S. and other available global sources to ensure comprehensive risk assessments.       

Tasks and Responsibilities :

•             Data Collection: Acquiring and preprocessing flood-related geospatial and climate data from the U.S. and other global sources.

•             Model Development: Implementing and optimizing Tab Transformer algorithms for flood susceptibility mapping.

•             Optimization: Applying meta-heuristic techniques (e.g., Genetic Algorithm, Particle Swarm Optimization) to improve the model's accuracy and performance.

•             Validation: Validating the model using real-world flood data and comparing its performance with traditional machine learning models.

•             Reporting: Preparing reports, high-resolution flood susceptibility maps, and presentations to communicate results.

Minimum Qualifications

•             Educational background in geo-Informatics Engineering.

•             Proficiency in machine learning, especially with deep learning models such as Tab Transformers, and familiarity with optimization algorithms.

•             Experience working with geospatial data and GIS software such as ArcGIS or QGIS.

•             Programming skills in Python, including libraries such as TensorFlow, Pandas, Numpy, Seaborn, and Scikit-learn.

•             Research experience in geospatial analysis and environmental hazard modeling, particularly in flood risk analysis

Terms of the Project: One Year

َ Mentor: ABOLGHASEM SADEGHI-NIARAKI              

Job Title: Associate Professor

Organization: Department of Computer Science and Engineering, Sejong University, South Korea         

Email: a.sadeghi@sejong.ac.kr       

Key Words: GeoAI, Flood Susceptibility Mapping, Tab Transformer Algorithms, Meta-Heuristic Optimization, Climate Change, Spatial Data Analysis      

Research Topics: Flood Susceptibility Mapping,Climate Change,Geospatial Artificial Intelligence (GeoAI)