Sawant, Shrutika S. and Erick, F. X. and St Goeb, and Holzer, Nina and Lang, Elmar W. and Goetz, Theresa (2023) An adaptive binary particle swarm optimization for solving multi-objective convolutional filter pruning problem. JOURNAL OF SUPERCOMPUTING, 79 (12). pp. 13287-13306. ISSN 0920-8542, 1573-0484
Full text not available from this repository.Abstract
In recent years, deep convolutional neural networks (DCNN) have evolved significantly in order to demonstrate remarkable performance in various computer vision tasks. However, their excess storage requirements and heavy computational burden restrict their scope of application, particularly on embedded platforms. This problem has motivated the research community to investigate effective approaches that can reduce computational burden without compromising its performance. Filter pruning is one of the popular ways to reduce the computational burden, where weak or unimportant convolutional filters are eliminated. In this paper, we propose a novel approach for filter pruning based on an adaptive multi-objective particle swarm optimization (AMPSO) to compress and accelerate DCNN. The proposed approach searches for an optimal solution while maintaining the trade-off between network's performance and computational cost. Extensive experiments on TernausNet and U-Net for high-resolution aerial image segmentation tasks demonstrate the superiority of AMPSO in finding a compact network model.
| Item Type: | Article |
|---|---|
| Uncontrolled Keywords: | NEURAL-NETWORKS; EFFICIENT; Deep convolutional neural network (DCNN); Filter pruning; Multi-objective optimization; Particle swarm optimization |
| Subjects: | 000 Computer science, information & general works > 004 Computer science 500 Science > 570 Life sciences |
| Divisions: | Biology, Preclinical Medicine > Institut für Biophysik und physikalische Biochemie > Prof. Dr. Elmar Lang |
| Depositing User: | Dr. Gernot Deinzer |
| Date Deposited: | 13 Mar 2024 14:23 |
| Last Modified: | 13 Mar 2024 14:23 |
| URI: | https://pred.uni-regensburg.de/id/eprint/59970 |
Actions (login required)
 |
View Item |
