Approximate Fisher information matrix to characterise the training of deep neural networks
Date
2020
Authors
Liao, Z.
Drummond, T.
Reid, I.
Carneiro, G.
Editors
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Journal article
Citation
IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020; 42(1):15-26
Statement of Responsibility
Zhibin Liao, Tom Drummond, Ian Reid, Gustavo Carneiro,
Conference Name
Abstract
In this paper, we introduce a novel methodology for characterising the performance of deep learning networks (ResNets and DenseNet) with respect to training convergence and generalisation as a function of mini-batch size and learning rate for image classification. This methodology is based on novel measurements derived from the eigenvalues of the approximate Fisher information matrix, which can be efficiently computed even for high capacity deep models. Our proposed measurements can help practitioners to monitor and control the training process (by actively tuning the mini-batch size and learning rate) to allow for good training convergence and generalisation. Furthermore, the proposed measurements also allow us to show that it is possible to optimise the training process with a new dynamic sampling training approach that continuously and automatically change the mini-batch size and learning rate during the training process. Finally, we show that the proposed dynamic sampling training approach has a faster training time and a competitive classification accuracy compared to the current state of the art.
School/Discipline
Dissertation Note
Provenance
Description
Date of Publication: 16 October 2018
Access Status
Rights
© IEEE