Difference between revisions of "Publications/kamal.22.xkdd"
From LRDE
| Line 4: | Line 4: | ||
| authors = Ataollah Kamal, Elouan Vincent, Marc Plantevit, Céline Robardet |
| authors = Ataollah Kamal, Elouan Vincent, Marc Plantevit, Céline Robardet |
||
| booktitle = ECML PKDD International Workshop on eXplainable Knowledge Discovery in Data Mining |
| booktitle = ECML PKDD International Workshop on eXplainable Knowledge Discovery in Data Mining |
||
| − | | title = Improving the |
+ | | title = Improving the Quality of Rule-Based GNN Explanations |
| address = Grenoble, France |
| address = Grenoble, France |
||
| abstract = Recent works have proposed to explain GNNs using activation rules. Activation rules allow to capture specific configurations in the embedding space of a given layer that is discriminant for the GNN decision. These rules also catch hidden features of input graphs. This requires to associate these rules to representative graphs. In this paper, we propose on the one hand an analysis of heuristic-based algorithms to extract the activation rulesand on the other hand the use of transport-based optimal graph distances to associate each rule with the most specific graph that triggers them. |
| abstract = Recent works have proposed to explain GNNs using activation rules. Activation rules allow to capture specific configurations in the embedding space of a given layer that is discriminant for the GNN decision. These rules also catch hidden features of input graphs. This requires to associate these rules to representative graphs. In this paper, we propose on the one hand an analysis of heuristic-based algorithms to extract the activation rulesand on the other hand the use of transport-based optimal graph distances to associate each rule with the most specific graph that triggers them. |
||
| Line 19: | Line 19: | ||
booktitle = <nowiki>{</nowiki>ECML PKDD International Workshop on eXplainable Knowledge |
booktitle = <nowiki>{</nowiki>ECML PKDD International Workshop on eXplainable Knowledge |
||
Discovery in Data Mining<nowiki>}</nowiki>, |
Discovery in Data Mining<nowiki>}</nowiki>, |
||
| − | title = <nowiki>{</nowiki>Improving the |
+ | title = <nowiki>{</nowiki>Improving the Quality of Rule-Based <nowiki>{</nowiki>GNN<nowiki>}</nowiki> Explanations<nowiki>}</nowiki>, |
year = <nowiki>{</nowiki>2022<nowiki>}</nowiki>, |
year = <nowiki>{</nowiki>2022<nowiki>}</nowiki>, |
||
address = <nowiki>{</nowiki>Grenoble, France<nowiki>}</nowiki>, |
address = <nowiki>{</nowiki>Grenoble, France<nowiki>}</nowiki>, |
||
| Line 34: | Line 34: | ||
graph distances to associate each rule with the most |
graph distances to associate each rule with the most |
||
specific graph that triggers them.<nowiki>}</nowiki>, |
specific graph that triggers them.<nowiki>}</nowiki>, |
||
| − | pages = <nowiki>{</nowiki>1-16<nowiki>}</nowiki>, |
+ | pages = <nowiki>{</nowiki>1--16<nowiki>}</nowiki>, |
note = <nowiki>{</nowiki>accepted<nowiki>}</nowiki> |
note = <nowiki>{</nowiki>accepted<nowiki>}</nowiki> |
||
<nowiki>}</nowiki> |
<nowiki>}</nowiki> |
||
Revision as of 16:13, 14 December 2022
- Authors
- Ataollah Kamal, Elouan Vincent, Marc Plantevit, Céline Robardet
- Where
- ECML PKDD International Workshop on eXplainable Knowledge Discovery in Data Mining
- Place
- Grenoble, France
- Type
- inproceedings
- Keywords
- IA
- Date
- 2022-09-12
Abstract
Recent works have proposed to explain GNNs using activation rules. Activation rules allow to capture specific configurations in the embedding space of a given layer that is discriminant for the GNN decision. These rules also catch hidden features of input graphs. This requires to associate these rules to representative graphs. In this paper, we propose on the one hand an analysis of heuristic-based algorithms to extract the activation rulesand on the other hand the use of transport-based optimal graph distances to associate each rule with the most specific graph that triggers them.
Bibtex (lrde.bib)
@InProceedings{ kamal.22.xkdd,
author = {Ataollah Kamal and Elouan Vincent and Marc Plantevit and
C\'{e}line Robardet},
booktitle = {ECML PKDD International Workshop on eXplainable Knowledge
Discovery in Data Mining},
title = {Improving the Quality of Rule-Based {GNN} Explanations},
year = {2022},
address = {Grenoble, France},
month = sep,
abstract = {Recent works have proposed to explain GNNs using
activation rules. Activation rules allow to capture
specific configurations in the embedding space of a given
layer that is discriminant for the GNN decision. These
rules also catch hidden features of input graphs. This
requires to associate these rules to representative graphs.
In this paper, we propose on the one hand an analysis of
heuristic-based algorithms to extract the activation rules,
and on the other hand the use of transport-based optimal
graph distances to associate each rule with the most
specific graph that triggers them.},
pages = {1--16},
note = {accepted}
}