EPFL - Ecole Polytechnique Fédérale de Lausanne

Open Opportunities

Visual condition monitoring of Railway retaining walls with machine learning EPFL - Ecole Polytechnique Fédérale de Lausanne ENAC - Civil Engineering Section This project is part of a collaboration between the IMOS lab and Matterhorn Gotthard Bahn, a railway company operating in the Swiss Alps. The student will work on developing computer vision algorithms for automated visual inspection of retaining walls around railway tracks. Retaining (or supporting) walls are crucial infrastructure elements responsible for maintaining the structural integrity of terrains around railway tracks and ensure safe operation. They are subject to wear and damages including cracks, concrete cancer (i.e., alkali–silica reaction), displacements, erosion and water infiltration. Images of retaining walls have already been collected and labels are available. The goal will be to design algorithms to estimate the condition of a wall, with a focus on robustness, transfer learning, and explainability (XAI). Engineering and Technology, Information, Computing and Communication Sciences Internship, Master Thesis, Semester Project

Multi-variate Time Series Anomaly Detection with Graph Neural Networks EPFL - Ecole Polytechnique Fédérale de Lausanne ENAC - Civil Engineering Section Predictive maintenance is critical in complex industrial systems as it aims to reduce maintenance costs and prevent unplanned downtime by detecting anomalies before they lead to system failure. Anomaly detection is a central component of predictive maintenance and involves identifying deviations from normal system behavior. However, constructing the graph from sensor recording data is often not trivial, because it is not clear how an edge between two time series should be defined. Transforming raw data into a graph representation is one of the key challenges faced in graph-based time series anomaly detection tasks. Time series data is often used to monitor the health of machines and equipment, but traditional approaches to predictive maintenance rely on statistical methods or machine learning models that do not capture the complex relationships between different data points. Graph Neural Networks (GNNs) have emerged as a promising approach for anomaly detection in such cases. GNNs encode graph structures by aggregating features of neighboring nodes, which allows them to capture both local and global structures of the graphs. By creating a graph structure from time series data, where each time series is represented as a node and edges capture interactions between the different time series, GNNs can analyze these relationships over time. In addition, GNNs can incorporate additional information about the time series as node and edge attributes, enabling them to capture nonlinear, large-scale, and heterogeneous relationships between time series. In addition, graphs can reveal fundamental insights into the system, which is valuable by itself, independent of an improvement in prediction quality. Furthermore, complex industrial systems are monitored by a variety of different sensors, which collect large amounts of heterogeneous data. The increasing availability and diversity of data resources offer new possibilities for more efficient and reliable system condition monitoring. However, the heterogeneity of sensor data presents several challenges in graph construction and learning from sensory time series. Artificial Intelligence and Signal and Image Processing, Engineering and Technology Bachelor Thesis, Master Thesis, Semester Project

Physics-infromed deep learning for predicting remaining load capacity of steel structures EPFL - Ecole Polytechnique Fédérale de Lausanne ENAC - Environmental Science and Engineering Section The project aims to develop computational models and physical experimentation methods to understand and predict the behavior of partially damaged steel members, enabling their reliable reuse rather than recycling. By creating a framework incorporating physics-informed deep learning methods, this prject aims to facilitate the incorporation of salvaged steel into future building designs, contributing significantly to a circular economy. Artificial Intelligence and Signal and Image Processing, Civil Engineering Master Thesis

Master thesis + Industrial Intership: Unleash Machine Learning Potential in Graph Neural Networks for Advanced Bridge Structural Health Monitoring EPFL - Ecole Polytechnique Fédérale de Lausanne ENAC - Civil Engineering Section ***Exciting Opportunity*** Are you intrigued by the potential of machine learning in the realm of structural engineering? Are you ready to make a real-world impact by utilizing AI to ensure the safety and reliability of our infrastructure? Launching in the Fall Semester of 2023 at EPFL, this Master's thesis offers the chance to apply advanced machine learning techniques, with a focus on Graph Neural Networks (GNNs), to the pivotal field of Bridge Structural Health Monitoring (SHM). ***Why This Matters*** Bridge SHM is a pivotal field designed to monitor the condition and performance of bridges continuously, ensuring their safety and reliability. Advanced technologies and data analysis methods, especially GNNs, are becoming increasingly vital in interpreting complex patterns from large datasets generated from diverse sensors. However, the heterogeneity of sensor data and capturing the temporal dynamics of these data points are significant challenges. Your work in this thesis can play a crucial role in overcoming these challenges, opening up new frontiers in the application of machine learning to structural health monitoring. Artificial Intelligence and Signal and Image Processing, Civil Engineering, Interdisciplinary Engineering Collaboration, Internship, Master Thesis, Semester Project

3D density map embedding for cryoEM EPFL - Ecole Polytechnique Fédérale de Lausanne Interfaculty Institute of Bioengineering (STI/SV) Cryo-Electron Microscopy (cryoEM) has opened unprecedented vistas in biology, offering detailed insights into macromolecular architectures. CryoEM captures snapshots of large complexes in their native configurations, enabling scientists to visualise their three-dimensional structures with unprecedented precision. However, the raw data is inherently noisy, and converting these 2D images into accurate 3D reconstructions demands meticulous postprocessing and advanced computational techniques. However, the challenge persists in dealing with various global and local resolutions and deriving precise structures from the cryoEM density maps. Owing to the inherently dynamic nature of proteins and their potential interactions with ligands and other subunits within complex systems, the task of effectively juxtaposing experimental cryoEM data with established reference structures presents an ongoing challenge. This project aims to construct density map embeddings encapsulating the information in the voxel grid through an autoencoder and evaluate their biological meaningfulness. Artificial Intelligence and Signal and Image Processing, Computational Biology and Bioinformatics, Structural Biology Collaboration, Master Thesis, Semester Project