2020
DOI: 10.1007/978-3-030-38040-3_52
|View full text |Cite
|
Sign up to set email alerts
|

Ablation of Artificial Neural Networks

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
3
0

Year Published

2021
2021
2024
2024

Publication Types

Select...
4
1

Relationship

0
5

Authors

Journals

citations
Cited by 5 publications
(4 citation statements)
references
References 9 publications
0
3
0
Order By: Relevance
“…To explore whether different components of the proposed network are required to accurately predict pCR, we conducted a set of ablation experiments 29 by modifying the network structure. First, we removed half of the network and dynamic information change capture branch, while remained a multi-task network for simultaneous tumor segmentation and pCR prediction (denoted as single-MTN) (Supplementary Figure S3).…”
Section: Methodsmentioning
confidence: 99%
“…To explore whether different components of the proposed network are required to accurately predict pCR, we conducted a set of ablation experiments 29 by modifying the network structure. First, we removed half of the network and dynamic information change capture branch, while remained a multi-task network for simultaneous tumor segmentation and pCR prediction (denoted as single-MTN) (Supplementary Figure S3).…”
Section: Methodsmentioning
confidence: 99%
“…The selected pixels are flipped by setting their values to zeros for the new input image to compute the new CNN prediction result so as to understand CNN behavior. V6: Visual Neuron Ablation : A neuron ablation interface allows users to select CNN layers and visualize their neuron‐level relevance data. Then, they can choose individual neurons or groups of neurons for ablation study [LMM18, VKS20], where the new prediction result when the selected neurons are removed can be compared with the original CNN prediction.…”
Section: Vislrpdesigner Design Overviewmentioning
confidence: 99%
“…In this paper, we develop a visual designer, named as VisLRPDesigner, which helps domain experts and students efficiently design, debug, and compare LRP models. It further integrates two visual analytics functions based on the computed relevance for model validation including: (1) pixel flipping which flips input image pixels to check CNN output changes, and (2) neuron ablation which removes specific neurons to see how that affects performance [VKS20]. The main contributions of this work are as follows: We construct an integrated computational framework of different LRP rules.…”
Section: Introductionmentioning
confidence: 99%
“…Cashew nuts are analyzed according to size, shape, color, and flaws using high-resolution cameras and sophisticated image processing algorithms. Machine vision techniques are used in [6] to assess the areca nut quality. For the grading process, six geometric features, three color features, and the fault area were taken into account.…”
Section: Introductionmentioning
confidence: 99%