In recent years, deep learning based methods have demonstrated state-of-the-art results on a number of different tasks. The key to deep learning success lies in its ability to handle large amounts of complex data and automatically extract a representation that is beneficial with respect to the task at hand. At the current state, deep learning underpins a wide range of technologies regarded as staples of modern day world, like facial recognition, natural language processing and autonomous veichles to name a few.
Deep learning is developing at a rapid pace, and the UiT Machine Learning Group has contributed several advances to this development. Our group has focused on developing novel methodology for unsupervised deep learning and forwarding deep learning methods for semantic segmentation. In particular, by utilizing aspects from kernel methods we have developed a novel deep learning approach for unsupervised learning.
1. Rethinking Knowledge Graph Propagation for Zero-Shot Learning (CVPR, 2019)
- Michael Kampffmeyer, Yinbo Chen, Xiaodan Liang , Hao Wang, Yujia Zhang, and Eric P. Xing
2. Understanding Convolutional Neural Networks With Information Theory: An Initial Exploration (IEEE Transactions on Neural Networks and Learning Systems, 2020)
- Shujian Yu, Kristoffer Wickstrøm, Robert Jenssen and José C. Príncipe