PhD candidate or Postdoctoral researcher in Machine Learning
The Qualcomm - UvA lab, established as part of the Informatics Institute, invites applications for a PhD-position in novel data efficient deep learning architectures.
Convolutional neural networks have revolutionized fields such as speech recognition and computer vision. Convolutions are efficient because they share parameters and implement equivariance with respect to translation, implying the same filters can be used anywhere in an image to detect an object. Equivariance with respect to translations can be generalized to other groups, such as rotations leading to even more parameter sharing and thus data efficiency*. The PhD candidate or postdoctoral researcher will study extensions of this idea, namely to exploit symmetries and other prior knowledge to design more data efficient convolutional networks.
What is requested for the project is a good understanding of deep neural networks and computer vision. A local, national and international training program will be part of the PhD. The candidate will be part of the QUVA-lab with the target is to perform the scientific research published in the best conferences, and to apply for patents awarded in the patent program.
* T.S. Cohen, M. Welling, Steerable CNNs. International Conference on Learning Representations (ICLR), 2017
This job comes from a partnership with Science Magazine and