Publications ‒ IMOS ‐ EPFL

At IMOS, our research encompasses both methodological contributions and practical applications in Prognostics and Health Management. Our methodological advancements span across diverse fields such as PHM, computer vision, and signal processing, while our applications address various domains, including energy and transportation. Through our work, we strive to bridge the gap between theoretical developments and real-world challenges.

Methodological Contributions

Practical Applications

List of all Publications

Algorithm-informed graph neural networks for leakage detection and localization in water distribution networks

Z. Zhang; O. Fink

Reliability Engineering and System Safety. 2026. Vol. 265, p. 111494. DOI : 10.1016/j.ress.2025.111494.

Methodological Contributions

Practical Applications

List of all Publications

Algorithm-informed graph neural networks for leakage detection and localization in water distribution networks

Interpretable prognostics with concept bottleneck models

Classifier-free diffusion-based weakly-supervised approach for health indicator derivation in rotating machines: Advancing early fault detection and condition monitoring

Equi-Euler GraphNet: An equivariant, temporal-dynamics informed graph neural network for dual force and trajectory prediction in multi-body systems

Explainable AI guided unsupervised fault diagnostics for high-voltage circuit breakers

Weighted Sum-Rate Maximization for Beamforming Design Using Minorization-Maximization: Convergence Rate and Deep Unfolding

Combining physics-based and data-driven modeling for building energy systems

Integrating Physics and Topology in Neural Networks for Learning Rigid Body Dynamics

Prescribing optimal health-aware operation for urban air mobility with deep reinforcement learning

GRADE: Generating Realistic and Dynamic Environments for robotics research with Isaac Sim

Automated processing of eXplainable Artificial Intelligence outputs in deep learning models for fault diagnostics of large infrastructures

Graph neural networks for virtual sensing in complex systems: Addressing heterogeneous temporal dynamics

ThermoNeRF: A multimodal Neural Radiance Field for joint RGB-thermal novel view synthesis of building facades

Interactive symbolic regression with co-design mechanism through offline reinforcement learning

Domain Adaptive Unfolded Graph Neural Networks

Power for AI and AI for Power: The Infinite Entanglement Between Artificial Intelligence and Power Electronics Systems

Exploiting semantic scene reconstruction for estimating building envelope characteristics

Overcoming Distribution Shifts and Imbalance Challenges in Representation Learning for Deep Regression Models

K-space physics-informed neural network (k-PINN) for compressed spectral mapping and efficient inversion of vibrations in thin composite laminates

Simplifying Source-Free Domain Adaptation for Object Detection: Effective Self-training Strategies and Performance Insights