Pruning by explaining: A novel criterion for deep neural network pruning
Seul-Ki Yeoma b, Philipp Seegerera c, Sebastian Lapuschkind, Alexander Bindere f, Simon M. Wiedemannd, Klaus‐Robert Müllera g h i✉, Wojciech Samekd g✉
- aMachine Learning Group, Technische Universität Berlin, 10587 Berlin, Germany
- bNota AI GmbH, 10117 Berlin, Germany
- cAignostics GmbH, 10557 Berlin, Germany
- dDepartment of Artificial Intelligence, Fraunhofer Heinrich Hertz Institute, 10587 Berlin, Germany
- eDepartment of Informatics, University of Oslo, 0373 Oslo, Norway
- fISTD Pillar, Singapore University of Technology and Design, Singapore 487372, Singapore
- gBIFOLD – Berlin Institute for the Foundations of Learning and Data, Berlin, Germany
- hDepartment of Artificial Intelligence, Korea University, Seoul 136–713, Korea
- iMax Planck Institut für Informatik, 66123 Saarbrücken, Germany
- ✉Corresponding author
Papers open inside SciRah from the original publisher or repository; SciRah never hosts them.
Abstract
The success of convolutional neural networks (CNNs) in various applications is accompanied by a significant increase in computation and parameter storage costs. Recent efforts to reduce these overheads involve pruning and compressing the weights of various layers while at the same time aiming to not sacrifice performance. In this paper, we propose a novel criterion for CNN pruning inspired by neural network interpretability: The most relevant units, i.e. weights or filters, are automatically found using their relevance scores obtained from concepts of explainable AI (XAI). By exploring this idea, we connect the lines of interpretability and model compression research. We show that our proposed method can efficiently prune CNN models in transfer-learning setups in which networks pre-trained on large corpora are adapted to specialized tasks. The method is evaluated on a broad range of computer vision datasets. Notably, our novel criterion is not only competitive or better compared to state-of-the-art pruning criteria when successive retraining is performed, but clearly outperforms these previous criteria in the resource-constrained application scenario in which the data of the task to be transferred to is very scarce and one chooses to refrain from fine-tuning. Our method is able to compress the model iteratively while maintaining or even improving accuracy. At the same time, it has a computational cost in the order of gradient computation and is comparatively simple to apply without the need for tuning hyperparameters for pruning.