3D Mesh Generative Models for Shape Optimization


Many practical continuous minimization problems, such as aerodynamic optimization, are not amenable to gradient-based optimization methods because derivatives can not be computed directly. We recently showed that it is possible to train a Neural Network regressor as a proxy to the numerical simulator and optimise the proxy function via Gradient-Descent. [ICML 2018 Paper]. However, in practice, we use a fixed topology, which is progressively deformed during the optimization process. In this project, we will develop a radically new approach to shape design for optimal aerodynamic by adjusting the level of details and let the Neural Network discover new topologies. This will let us explore much larger swathes of the shape space which may result in design solutions both different from human-conceived ones and objectively better.

Heat exchanger with complex topology
Starting point

The student will be implementing his work on top of a large state-of-the-art code base, which includes a Deep-Learning library for 3D shapes in the TensorFlow framework.

Key Domains
  • 3D Deep-Learning
  • Generative Models
  • Meshes
  • Proficiency in Python.
  • Knowledge of machine learning, deep-learning.
  • Basic knowledge in Computer Graphics is a plus.

Please send an email to pierre.baque@epfl.ch or timur.bagautdinov@epfl.ch

The work will be done in collaboration with the EPFL startup Neural Concept .