Sep 27, 2018 · In this paper we suggest trading accuracy with computation using a gated version of Convolutional Neural Networks (CNN). The gated network ...
The network's 'gates' imply which feature-maps are necessary for the task, and which are not. Specifically, full feature maps are considered for omission, to ...
Nov 25, 2019 · We propose to prune a neural network for a given single input to keep only neurons that highly contribute to the prediction.
Missing: Dependably | Show results with:Dependably
In view of the significant advantages of filter pruning, this paper proposes a graphical filter pruning strategy based on the similarity between feature maps.
Apr 3, 2023 · To solve the problem of complex network models with a large number of redundant parameters, a pruning algorithm combined with an attention mechanism is ...
We aim to approach the structural model pruning problem from a novel per- spective by exploiting the model's interpretations (a subset of input features called ...
Filter pruning (FP) is an effective method for reducing the computational costs of convolutional neural networks, and herein, the most critical task ...
Oct 2, 2021 · In this paper, two novel pruning methods are proposed. First, a new pruning method is proposed, which reflects the importance of filters by exploring the ...
Missing: Dependably | Show results with:Dependably
Input-Dependably Feature-Map Pruning: 27th International Conference on Artificial Neural Networks, Rhodes, Greece, October 4-7, 2018, Proceedings, Part I.
Jul 21, 2023 · In this work, we propose Adversarial Feature Map Pruning for Backdoor (FMP) to mitigate backdoor from the DNN.
Missing: Dependably | Show results with:Dependably