Why is unsupervised alignment of English embeddings from different algorithms so hard?

Mareike Hartmann, Yova Kementchedjhieva, Anders Søgaard. Why is unsupervised alignment of English embeddings from different algorithms so hard?. In Ellen Riloff, David Chiang 0001, Julia Hockenmaier, Jun'ichi Tsujii, editors, Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussels, Belgium, October 31 - November 4, 2018. pages 582-586, Association for Computational Linguistics, 2018. [doi]

@inproceedings{HartmannKS18,
  title = {Why is unsupervised alignment of English embeddings from different algorithms so hard?},
  author = {Mareike Hartmann and Yova Kementchedjhieva and Anders Søgaard},
  year = {2018},
  url = {https://aclanthology.info/papers/D18-1056/d18-1056},
  researchr = {https://researchr.org/publication/HartmannKS18},
  cites = {0},
  citedby = {0},
  pages = {582-586},
  booktitle = {Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, Brussels, Belgium, October 31 - November 4, 2018},
  editor = {Ellen Riloff and David Chiang 0001 and Julia Hockenmaier and Jun'ichi Tsujii},
  publisher = {Association for Computational Linguistics},
  isbn = {978-1-948087-84-1},
}