Learning multiple views with orthogonal denoising autoencoders
Ye, TengQi, Wang, Tianchun, McGuinness, KevinORCID: 0000-0003-1336-6477, Guo, Yu and Gurrin, CathalORCID: 0000-0003-2903-3968
(2016)
Learning multiple views with orthogonal denoising autoencoders.
In: The 22nd International Conference on Multimedia Modelling (MMM'16), 4-6 Jan 2016, Miami, FA..
ISBN 978-3-319-27673-1
Multi-view learning techniques are necessary when data is
described by multiple distinct feature sets because single-view learning algorithms tend to overt on these high-dimensional data. Prior successful approaches followed either consensus or complementary principles. Recent work has focused on learning both the shared and private latent spaces of views in order to take advantage of both principles. However, these methods can not ensure that the latent spaces are strictly independent through encouraging the orthogonality in their objective functions. Also little work has explored representation learning techniques for multiview learning. In this paper, we use the denoising autoencoder to learn shared and private latent spaces, with orthogonal constraints | disconnecting every private latent space from the remaining views. Instead of computationally expensive optimization, we adapt the backpropagation algorithm to train our model.
Proceedings of MMM 2016 - The 22nd International Conference on Multimedia Modeling. Lecture Notes in Computer Science
9517.
Springer. ISBN 978-3-319-27673-1