t-Distributed Stochastic Neighbor Embedding with Inhomogeneous Degrees of Freedom
One of the dimension reduction (DR) methods for data-visualization, t-distributed stochastic neighbor embedding (t-SNE), has drawn increasing attention. t-SNE gives us better visualization than conventional DR methods, by relieving so-called crowding problem. The crowding problem is one of the curses of dimensionality, which is caused by discrepancy between high and low dimensional spaces. However, in t-SNE, it is assumed that the strength of the discrepancy is the same for all samples in all datasets regardless of ununiformity of distributions or the difference in dimensions, and this assumption sometimes ruins visualization. Here we propose a new DR method inhomogeneous t-SNE, in which the strength is estimated for each point and dataset. Experimental results show that such pointwise estimation is important for reasonable visualization and that the proposed method achieves better visualization than the original t-SNE.
KeywordsSNE t-SNE Dimensionality reduction Degrees of freedom
This work was partially supported by the Grant for Enhancement of International Research from Kobe University, and Grants-in-Aid for Young Scientists (B) [No. 15K16064 (J.K.)] from the MEXT of Japan.
- 4.Hinton, G.E., Roweis, S.T.: Stochastic neighbor embedding. In: Advances in Neural Information Processing Systems, vol. 15, pp. 833–840. MIT Press, Cambridge (2002)Google Scholar
- 7.van der Maaten, L.: Learning a parametric embedding by preserving local structure. In: International Conference on Artificial Intelligence and Statistics, JMLR W&CP, vol. 5 (2009)Google Scholar
- 8.Vladymyrov, M., Carreira-Perpinán, M.: Entropic affinities: properties and efficient numerical computation. In: Proceedings of the 30th International Conference on Machine Learning, pp. 477–485 (2013)Google Scholar
- 9.Vladymyrov, M., Carreira-Perpinán, M.A.: Linear-time training of nonlinear low-dimensional embeddings. In: Proceedings of AISTATS 2014, International Conference on Artificial Intelligence and Statistics, JMLR W&CP, vol. 33, pp. 968–977 (2014)Google Scholar
- 10.van der Maaten, L.: Barnes-hut-sne. arXiv preprint arXiv:1301.3342 (2013)