- LEVEL ACCURACY WITH 50 X FEWER PARAMETERS AND < 0 . 5 MB MODEL SIZE

By Forrest N. Iandola, Song Han, Matthew W. Moskewicz, Khalid Ashraf, William (Bill) J. Dally, Kurt Keutzer. 2016

Subfields: Computer Vision and Pattern Recognition, Efficiency & Compression

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Efficiency & Compression

Citations in the corpus

Similarity-Preserving Knowledge Distillation (2019, background): “Resource efficiency considerations have led to a recent increase in interest in efficient neural architectures [13, 14, 23, 32, 44], as well as in algorithms for compressing trained deep networks.”
N2N Learning: Network to Network Compression via Policy Gradient Reinforcement Learning (2017, background): “…2006), (Miikkulainen et al. 2017), (Real et al. 2017), (Snoek, Larochelle, and Adams 2012), (Snoek et al. 2015), (Stanley and Miikkulainen 2002), (Jozefowicz, Zaremba, and Sutskever 2015), (Murdock et al. 2016), (Feng and Darrell 2015), (Warde-Farley, Rabinovich, and Anguelov 2014), (Iandola et…”
Be Your Own Teacher: Improve the Performance of Convolutional Neural Networks via Self Distillation (2019, methodology): “Such techniques include lightweight networks design [19, 16], pruning [12, 13] and quantization [5, 31].”
EnergyNet: Energy-Efficient Dynamic Inference (2018, background): “Most of them aim to reduce the model size or the number of 28 computations [3, 4, 5, 6, 7, 8, 9, 10, 11, 12].”
DeepAnT: A Deep Learning Approach for Unsupervised Anomaly Detection in Time Series (2019, methodology): “In DeepAnt, similar to other well-known methods [44], [45], the output of a convolutional layer is fur-”
Energy-efficient Amortized Inference with Cascaded Deep Classifiers (2017, methodology): “Other techniques include pruning of connections [Han et al., 2015] and bottleneck structure [Iandola et al., 2016].”