Disjoint Label Space Transfer Learning with Common Factorised Space


  • Xiaobin Chang Queen Mary University of London
  • Yongxin Yang University of Edinburgh
  • Tao Xiang Queen Mary University of London
  • Timothy M. Hospedales Edinburgh University




In this paper, a unified approach is presented to transfer learning that addresses several source and target domain labelspace and annotation assumptions with a single model. It is particularly effective in handling a challenging case, where source and target label-spaces are disjoint, and outperforms alternatives in both unsupervised and semi-supervised settings. The key ingredient is a common representation termed Common Factorised Space. It is shared between source and target domains, and trained with an unsupervised factorisation loss and a graph-based loss. With a wide range of experiments, we demonstrate the flexibility, relevance and efficacy of our method, both in the challenging cases with disjoint label spaces, and in the more conventional cases such as unsupervised domain adaptation, where the source and target domains share the same label-sets.




How to Cite

Chang, X., Yang, Y., Xiang, T., & Hospedales, T. M. (2019). Disjoint Label Space Transfer Learning with Common Factorised Space. Proceedings of the AAAI Conference on Artificial Intelligence, 33(01), 3288-3295. https://doi.org/10.1609/aaai.v33i01.33013288



AAAI Technical Track: Machine Learning