About me
Hello, and thanks for stopping by!
My name is Keivan Faghih Niresi (Persian: کیوان فقیه نیرسی). Since February 2023, I have been pursuing a Ph.D. at the Intelligent Maintenance and Operations Systems (IMOS) Laboratory, École Polytechnique Fédérale de Lausanne (EPFL) in Switzerland. My doctoral thesis focuses on reliable data-driven computational sensing methods through signal processing and machine learning on graphs, under the supervision of Prof. Olga Fink.
Prior to joining EPFL, I earned my master's degree from the Institute of Communications Engineering, College of Electrical Engineering and Computer Science, National Tsing Hua University (NTHU) in Taiwan, where I conducted research in convex and non-convex optimization, robust statistics, deep learning, and hyperspectral imaging at the Wireless Communications and Signal Processing (WCSP) Laboratory under the supervision of Prof. Chong-Yung Chi. I also had the opportunity to work as a machine learning engineer intern at PranaQ, where I focused on developing mathematical algorithms for multimodal biomedical data.
The foundation of my research lies in advancing computational sensing and imaging through the development of innovative algorithms and mathematical frameworks. I am particularly interested in topics at the intersection of machine learning, signal processing, and computational mathematics, such as inverse problems, graph representation learning, domain adaptation, physics-informed learning, and optimization. These methods are applied to diverse applications, including low-level vision (image restoration, fusion, compression, etc.), Internet of things, remote sensing, intelligent infrastructures, and smart cities. My research emphasizes two key aspects: 1) exploring the fundamental mathematical principles of sensing, and 2) pursuing application-driven projects through collaborations with experts from various disciplines such as electrical engineering, computer science, civil and environmental engineering, applied mathematics, mechanical engineering, and physics.
Featured Publications
![]() | Physics-Enhanced Graph Neural Networks for Soft Sensing in Industrial Internet of Things Keivan Faghih Niresi, Hugo Bissig, Henri Baumann, and Olga Fink IEEE Internet of Things Journal, 2024 IEEE Xplore | PDF | Code |
---|---|
![]() | Robust Hyperspectral Inpainting via Low-Rank Regularized Untrained Convolutional Neural Network Keivan Faghih Niresi, and Chong-Yung ChiIEEE Geoscience and Remote Sensing Letters, 2023 IEEE Xplore | PDF | Code |
![]() | Unsupervised Hyperspectral Denoising Based on Deep Image Prior and Least Favorable Distribution Keivan Faghih Niresi, and Chong-Yung ChiIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2022 IEEE Xplore | PDF | Code | Open Remote Sensing |
News
Our research paper, "Physics-Enhanced Graph Neural Networks For Soft Sensing in Industrial Internet of Things", has been accepted by IEEE Internet of Things Journal!
Our research paper, "Informed Graph Learning By Domain Knowledge Injection and Smooth Graph Signal Representation", has been accepted by EUSIPCO 2024. See you in Lyon!