Relevance Propagation through Deep Conditional Random Fields

Onderzoeksoutput: Conference paper

Samenvatting

Conditional random fields (CRFs), a particular type of graph neural networks (GNNs), can be used to make structured predictions in machine learning, with various applications from image processing and natural language processing to recommender systems. CRFs refine the prediction of a sample by taking into account its context information. However, there is a lack of work on post-hoc explanation approaches to CRFs, especially when the model is softmax-activated like the deep mean field network (DMFN). In this paper, we bridge this gap by proposing a layer-wise relevance propagation (LRP) method based on deep Taylor decomposition to explain CRFs, especially the DMFN model. The method considers the intermediate softmax activation layers in DMFN. We use two evaluation settings: top K\% deletion and insertion to evaluate the method. Experimental studies on fake news detection using the DMFN model prove the effectiveness of our explanation method compared to the other baseline methods.
Originele taal-2English
TitelICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing
UitgeverijIEEE
Pagina's1-5
Aantal pagina's5
ISBN van elektronische versie978-1-7281-6327-7
DOI's
StatusPublished - 5 mei 2023
Evenement2023 IEEE International Conference on Acoustics, Speech, and Signal Processing - Rodos Palace Luxury Convention Resort, Rhodes, Greece
Duur: 4 jun 202310 jun 2023
https://2023.ieeeicassp.org/

Publicatie series

NaamICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume2023-June
ISSN van geprinte versie1520-6149

Conference

Conference2023 IEEE International Conference on Acoustics, Speech, and Signal Processing
Verkorte titelICASSP 2023
Land/RegioGreece
StadRhodes
Periode4/06/2310/06/23
Internet adres

Bibliografische nota

Publisher Copyright:
© 2023 IEEE.

Vingerafdruk

Duik in de onderzoeksthema's van 'Relevance Propagation through Deep Conditional Random Fields'. Samen vormen ze een unieke vingerafdruk.

Citeer dit