Deep learning and transfer learning

We work across a number of deep learning paradigms. We're leaders in the automated design of deep learning models, including deep neural networks.

Light scribbles coalesce in shape of globe against dark background

Deep learning paradigms

We work on a number of different deep learning paradigms, including:

  • neural network based deep learning such as deep convolutional neural networks
  • deep auto-encoders
  • deep generative adversarial networks, deep belief networks, and their variants
  • non-neural network-based deep learning such as:
    • GP-based deep learning
    • deep PCAs
    • deep forest learning.

Automated design

Te Herenga Waka—Victoria University of Wellington is particularly good in automated design of deep learning models.

This includes deep neural networks and other deep models, where the structure or architecture and parameters are learned and optimised automatically and simultaneously.

Transfer learning

Transfer learning is also a big area of the Centre, including:

  • domain adaptation
  • domain generalisation
  • multi-task learning and optimisation.

Learning paradigms

The work varies with different learning paradigms such as:

  • evolutionary computation techniques
  • neural networks
  • kerne-based learning or symbolic learning algorithms.