Abstract | ||
---|---|---|
Model selection methods based on stochastic regularization have been widely used in deep learning due to their simplicity and effectiveness. The well-known Dropout method treats all units, visible or hidden, in the same way, thus ignoring any a priori information related to grouping or structure. Such structure is present in multi-modal learning applications such as affect analysis and gesture recognition, where subsets of units may correspond to individual modalities. Here we describe Modout, a model selection method based on stochastic regularization, which is particularly useful in the multi-modal setting. Different from other forms of stochastic regularization, it is capable of learning whether or when to fuse two modalities in a layer, which is usually considered to be an architectural hyper-parameter by deep learning researchers and practitioners. Modout is evaluated on two real multi-modal datasets. The results indicate improved performance compared to other forms of stochastic regularization. The result on the Montalbano dataset shows that learning a fusion structure by Modout is on par with a state-of-the-art carefully designed architecture. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/FG.2017.59 | 2017 12th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2017) |
Keywords | Field | DocType |
Modout,multimodal architecture learning,stochastic regularization,model selection,deep learning,Dropout method,Montalbano dataset,fusion structure learning | Semi-supervised learning,A priori and a posteriori,Gesture recognition,Model selection,Regularization (mathematics),Artificial intelligence,Deep learning,Mathematics,Machine learning,Modal,Regularization perspectives on support vector machines | Conference |
ISSN | ISBN | Citations |
2326-5396 | 978-1-5090-4024-7 | 3 |
PageRank | References | Authors |
0.38 | 25 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Fan Li | 1 | 40 | 4.95 |
Natalia Neverova | 2 | 265 | 14.44 |
Christian Wolf | 3 | 1027 | 54.93 |
Graham W. Taylor | 4 | 1523 | 127.22 |