Postdoctoral Researcher in Generative Modelling at DTU, Denmark (2-year)
299 views
Skip to first unread message
frel...@gmail.com
May 17, 2024, 6:10:07 AM5/17/24
to Machine Learning News
Summary:
We are seeking a postdoctoral researcher at DTU specialising in generative modelling. This position involves developing and advancing methodological techniques, particularly diffusion and score-based models, and applying these techniques to physical systems. The ideal candidate will hold a Ph.D. in machine learning, have substantial experience with deep generative models, and possess strong programming skills.
Position Overview:
The 2-year postdoc position at the Technical University of Denmark (DTU) focuses on integrating physics-aware machine learning to enhance the computational efficiency of simulations critical for understanding natural phenomena and advancing technologies for environmental sustainability. We aim to enable larger and more complex simulations of physical systems by solving the physics only at a few discrete points and then using deep generative models to fill in the remaining space accurately. Initially, this approach will be applied to magnetic fields due to their significant role in sustainable applications and well-understood physical properties.
The role emphasises methodological advancements in deep generative models, particularly diffusion and score-based models. The primary task will be to develop generative models capable of interpolating and extrapolating in a physically correct manner, ensuring that the models adhere to Maxwell’s equations governing magnetic fields. This methodology will later be adapted to other physical systems by defining appropriate models for each. The postdoc will be supervised by Associate Professor Jes Frellsen at DTU Compute and co-supervised by Professor Rasmus Bjørk from DTU Energy.
Qualifications:
The ideal candidate is expected to have a strong machine-learning background and hold a PhD in machine learning or a closely related field. Experience in deep generative model research is essential. International experience will be considered advantageous. Proficiency in programming with Python, PyTorch, TensorFlow, or Jax, as well as a good command of the English language, are essential.
Contact Information:
If you are interested in this position or require further information, please contact Associate Professor Jes Frellsen (https://frellsen.org).
We are seeking a postdoctoral researcher at DTU specialising in generative modelling. This position involves developing and advancing methodological techniques, particularly diffusion and score-based models, and applying these techniques to physical systems. The ideal candidate will hold a Ph.D. in machine learning, have substantial experience with deep generative models, and possess strong programming skills.
Position Overview:
The 2-year postdoc position at the Technical University of Denmark (DTU) focuses on integrating physics-aware machine learning to enhance the computational efficiency of simulations critical for understanding natural phenomena and advancing technologies for environmental sustainability. We aim to enable larger and more complex simulations of physical systems by solving the physics only at a few discrete points and then using deep generative models to fill in the remaining space accurately. Initially, this approach will be applied to magnetic fields due to their significant role in sustainable applications and well-understood physical properties.
The role emphasises methodological advancements in deep generative models, particularly diffusion and score-based models. The primary task will be to develop generative models capable of interpolating and extrapolating in a physically correct manner, ensuring that the models adhere to Maxwell’s equations governing magnetic fields. This methodology will later be adapted to other physical systems by defining appropriate models for each. The postdoc will be supervised by Associate Professor Jes Frellsen at DTU Compute and co-supervised by Professor Rasmus Bjørk from DTU Energy.
Qualifications:
The ideal candidate is expected to have a strong machine-learning background and hold a PhD in machine learning or a closely related field. Experience in deep generative model research is essential. International experience will be considered advantageous. Proficiency in programming with Python, PyTorch, TensorFlow, or Jax, as well as a good command of the English language, are essential.
Contact Information:
If you are interested in this position or require further information, please contact Associate Professor Jes Frellsen (https://frellsen.org).
Reply all
Reply to author
Forward
0 new messages