Lazy Evaluation of Convolutional Filters
Sam Leroux, Steven Bohez, Cedric De Boom, Elias De Coninck, Tim Verbelen, Bert Vankeirsbilck, Pieter Simoens, Bart Dhoedt
arXiv:1605.08543v17 citations
Originality Synthesis-oriented
AI Analysis
This addresses efficiency challenges for deployment on constrained devices, but appears incremental as it builds on existing convolutional network methods.
The paper tackles the problem of computational and memory constraints in deep neural networks by proposing a technique that avoids evaluating certain convolutional filters, enabling a trade-off between accuracy and resource requirements.
In this paper we propose a technique which avoids the evaluation of certain convolutional filters in a deep neural network. This allows to trade-off the accuracy of a deep neural network with the computational and memory requirements. This is especially important on a constrained device unable to hold all the weights of the network in memory.