Compression strategies and space-conscious representations for deep neural networks
Contributo in Atti di convegno
Data di Pubblicazione:
2020
Citazione:
Compression strategies and space-conscious representations for deep neural networks / G.C. Marino, G. Ghidoli, M. Frasca, D. Malchiodi (INTERNATIONAL CONFERENCE ON PATTERN RECOGNITION). - In: 2020 25th International Conference on Pattern Recognition (ICPR)[s.l] : IEEE, 2020. - ISBN 978-1-7281-8808-9. - pp. 9835-9842 (( Intervento presentato al 25. convegno International Conference on Pattern Recognition, ICPR 2020 tenutosi a Milano nel 2021 [10.1109/ICPR48806.2021.9412209].
Abstract:
Recent advances in deep learning have made available large, powerful convolutional neural networks (CNN) with state-of-the-art performance in several real-world applications. Unfortunately, these large-sized models have millions of parameters, thus they are not deployable on resource-limited platforms (e.g., where RAM is limited). Compression of CNNs becomes therefore a critical problem to achieve memory-efficient and possibly computationally faster model representations. In this paper, we investigate the impact of lossy compression of CNNs by weight pruning and quantization, and lossless weight matrix representations based on source coding. We tested several combinations of these techniques on four benchmark datasets for classification and regression problems, achieving compression rates up to 165 times, while preserving or improving the model performance.
Tipologia IRIS:
03 - Contributo in volume
Keywords:
CNN compression; Drug-target prediction; Entropy coding; Probabilistic quantization; Weight pruning
Elenco autori:
G.C. Marino, G. Ghidoli, M. Frasca, D. Malchiodi
Link alla scheda completa:
Titolo del libro:
2020 25th International Conference on Pattern Recognition (ICPR)