Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction
Abstract Background Deep Neural Networks (DNN), in particular, Convolutional Neural Networks (CNN), has recently achieved state-of-art results for the task of Drug-Drug Interaction (DDI) extraction. Most CNN architectures incorporate a pooling layer to reduce the dimensionality of the convolution la...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
BMC
2018-06-01
|
Series: | BMC Bioinformatics |
Subjects: | |
Online Access: | http://link.springer.com/article/10.1186/s12859-018-2195-1 |
id |
doaj-dba792d88a2843b1b309771fc836fc90 |
---|---|
record_format |
Article |
spelling |
doaj-dba792d88a2843b1b309771fc836fc902020-11-24T22:17:20ZengBMCBMC Bioinformatics1471-21052018-06-0119S8394710.1186/s12859-018-2195-1Evaluation of pooling operations in convolutional architectures for drug-drug interaction extractionVíctor Suárez-Paniagua0Isabel Segura-Bedmar1Computer Science Department, Carlos III University of MadridComputer Science Department, Carlos III University of MadridAbstract Background Deep Neural Networks (DNN), in particular, Convolutional Neural Networks (CNN), has recently achieved state-of-art results for the task of Drug-Drug Interaction (DDI) extraction. Most CNN architectures incorporate a pooling layer to reduce the dimensionality of the convolution layer output, preserving relevant features and removing irrelevant details. All the previous CNN based systems for DDI extraction used max-pooling layers. Results In this paper, we evaluate the performance of various pooling methods (in particular max-pooling, average-pooling and attentive pooling), as well as their combination, for the task of DDI extraction. Our experiments show that max-pooling exhibits a higher performance in F1-score (64.56%) than attentive pooling (59.92%) and than average-pooling (58.35%). Conclusions Max-pooling outperforms the others alternatives because is the only one which is invariant to the special pad tokens that are appending to the shorter sentences known as padding. Actually, the combination of max-pooling and attentive pooling does not improve the performance as compared with the single max-pooling technique.http://link.springer.com/article/10.1186/s12859-018-2195-1Deep learningConvolutional neural networkPoolingAttention modelDrug-drug interaction extraction |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Víctor Suárez-Paniagua Isabel Segura-Bedmar |
spellingShingle |
Víctor Suárez-Paniagua Isabel Segura-Bedmar Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction BMC Bioinformatics Deep learning Convolutional neural network Pooling Attention model Drug-drug interaction extraction |
author_facet |
Víctor Suárez-Paniagua Isabel Segura-Bedmar |
author_sort |
Víctor Suárez-Paniagua |
title |
Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction |
title_short |
Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction |
title_full |
Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction |
title_fullStr |
Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction |
title_full_unstemmed |
Evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction |
title_sort |
evaluation of pooling operations in convolutional architectures for drug-drug interaction extraction |
publisher |
BMC |
series |
BMC Bioinformatics |
issn |
1471-2105 |
publishDate |
2018-06-01 |
description |
Abstract Background Deep Neural Networks (DNN), in particular, Convolutional Neural Networks (CNN), has recently achieved state-of-art results for the task of Drug-Drug Interaction (DDI) extraction. Most CNN architectures incorporate a pooling layer to reduce the dimensionality of the convolution layer output, preserving relevant features and removing irrelevant details. All the previous CNN based systems for DDI extraction used max-pooling layers. Results In this paper, we evaluate the performance of various pooling methods (in particular max-pooling, average-pooling and attentive pooling), as well as their combination, for the task of DDI extraction. Our experiments show that max-pooling exhibits a higher performance in F1-score (64.56%) than attentive pooling (59.92%) and than average-pooling (58.35%). Conclusions Max-pooling outperforms the others alternatives because is the only one which is invariant to the special pad tokens that are appending to the shorter sentences known as padding. Actually, the combination of max-pooling and attentive pooling does not improve the performance as compared with the single max-pooling technique. |
topic |
Deep learning Convolutional neural network Pooling Attention model Drug-drug interaction extraction |
url |
http://link.springer.com/article/10.1186/s12859-018-2195-1 |
work_keys_str_mv |
AT victorsuarezpaniagua evaluationofpoolingoperationsinconvolutionalarchitecturesfordrugdruginteractionextraction AT isabelsegurabedmar evaluationofpoolingoperationsinconvolutionalarchitecturesfordrugdruginteractionextraction |
_version_ |
1725785382826541056 |