Matches in SemOpenAlex for { <https://semopenalex.org/work/W4383102502> ?p ?o ?g. }
- W4383102502 abstract "Graph Neural Networks (GNNs) have achieved promising performance in a variety of practical applications. Similar to traditional DNNs, GNNs could exhibit incorrect behavior that may lead to severe consequences, and thus testing is necessary and crucial. However, labeling all the test inputs for GNNs can be costly and time-consuming, especially when dealing with large and complex graphs, which seriously affects the efficiency of GNN testing. Existing studies have focused on test prioritization for DNNs, which aims to identify and prioritize fault-revealing tests (i.e., test inputs that are more likely to be misclassified) to detect system bugs earlier in a limited time. Although some DNN prioritization approaches have been demonstrated effective, there is a significant problem when applying them to GNNs: they do not take into account the connections (edges) between GNN test inputs (nodes), which play a significant role in GNN inference. In general, DNN test inputs are independent of each other, while GNN test inputs are usually represented as a graph with complex relationships between each test. In this paper, we propose GraphPrior ( GNN -oriented Test Prior itization), a set of approaches to prioritize test inputs specifically for GNNs via mutation analysis. Inspired by mutation testing in traditional software engineering, in which test suites are evaluated based on the mutants they kill, GraphPrior generates mutated models for GNNs and regards test inputs that kill many mutated models as more likely to be misclassified. Then, GraphPrior leverages the mutation results in two ways, killing-based and feature-based methods. When scoring a test input, the killing-based method considers each mutated model equally important, while feature-based methods learn different importance for each mutated model through ranking models. Finally, GraphPrior ranks all the test inputs based on their scores. We conducted an extensive study based on 604 subjects to evaluate GraphPrior on both natural and adversarial test inputs. The results demonstrate that KMGP, the killing-based GraphPrior approach, outperforms the compared approaches in a majority of cases, with an average improvement of 4.76%–49.60% in terms of APFD. Furthermore, the feature-based GraphPrior approach, RFGP, performs the best among all the GraphPrior approaches. On adversarial test inputs, RFGP outperforms the compared approaches across different adversarial attacks, with the average improvement of 2.95%–46.69%." @default.
- W4383102502 created "2023-07-05" @default.
- W4383102502 creator A5006693940 @default.
- W4383102502 creator A5040326968 @default.
- W4383102502 creator A5040574362 @default.
- W4383102502 creator A5059687220 @default.
- W4383102502 creator A5081145634 @default.
- W4383102502 creator A5082835974 @default.
- W4383102502 date "2023-07-04" @default.
- W4383102502 modified "2023-10-16" @default.
- W4383102502 title "GraphPrior: Mutation-based Test Input Prioritization for Graph Neural Networks" @default.
- W4383102502 cites W1967187747 @default.
- W4383102502 cites W1991292921 @default.
- W4383102502 cites W2014515160 @default.
- W4383102502 cites W2049695835 @default.
- W4383102502 cites W2073125568 @default.
- W4383102502 cites W2104942599 @default.
- W4383102502 cites W2110068396 @default.
- W4383102502 cites W2120563984 @default.
- W4383102502 cites W2126210726 @default.
- W4383102502 cites W2134512579 @default.
- W4383102502 cites W2163559687 @default.
- W4383102502 cites W2171451409 @default.
- W4383102502 cites W2240451833 @default.
- W4383102502 cites W2292289085 @default.
- W4383102502 cites W2392262895 @default.
- W4383102502 cites W2551441958 @default.
- W4383102502 cites W2604319603 @default.
- W4383102502 cites W2616028256 @default.
- W4383102502 cites W2616524285 @default.
- W4383102502 cites W2617809069 @default.
- W4383102502 cites W2725449579 @default.
- W4383102502 cites W2799640043 @default.
- W4383102502 cites W2803831897 @default.
- W4383102502 cites W2808957028 @default.
- W4383102502 cites W2905627582 @default.
- W4383102502 cites W2914721378 @default.
- W4383102502 cites W2922015121 @default.
- W4383102502 cites W2954629067 @default.
- W4383102502 cites W2962946486 @default.
- W4383102502 cites W2963161254 @default.
- W4383102502 cites W2963308851 @default.
- W4383102502 cites W2963857521 @default.
- W4383102502 cites W2966149470 @default.
- W4383102502 cites W2971567992 @default.
- W4383102502 cites W3041012898 @default.
- W4383102502 cites W3042703469 @default.
- W4383102502 cites W3045200674 @default.
- W4383102502 cites W3047443805 @default.
- W4383102502 cites W3095937012 @default.
- W4383102502 cites W3099444373 @default.
- W4383102502 cites W3100848837 @default.
- W4383102502 cites W3102476541 @default.
- W4383102502 cites W3105347387 @default.
- W4383102502 cites W3110933132 @default.
- W4383102502 cites W3120991880 @default.
- W4383102502 cites W3123909522 @default.
- W4383102502 cites W3128290931 @default.
- W4383102502 cites W3152893301 @default.
- W4383102502 cites W3161493619 @default.
- W4383102502 cites W3178469298 @default.
- W4383102502 cites W3214309058 @default.
- W4383102502 cites W4225410698 @default.
- W4383102502 cites W4226237846 @default.
- W4383102502 cites W4296746241 @default.
- W4383102502 doi "https://doi.org/10.1145/3607191" @default.
- W4383102502 hasPublicationYear "2023" @default.
- W4383102502 type Work @default.
- W4383102502 citedByCount "0" @default.
- W4383102502 crossrefType "journal-article" @default.
- W4383102502 hasAuthorship W4383102502A5006693940 @default.
- W4383102502 hasAuthorship W4383102502A5040326968 @default.
- W4383102502 hasAuthorship W4383102502A5040574362 @default.
- W4383102502 hasAuthorship W4383102502A5059687220 @default.
- W4383102502 hasAuthorship W4383102502A5081145634 @default.
- W4383102502 hasAuthorship W4383102502A5082835974 @default.
- W4383102502 hasBestOaLocation W43831025021 @default.
- W4383102502 hasConcept C104317684 @default.
- W4383102502 hasConcept C119857082 @default.
- W4383102502 hasConcept C124101348 @default.
- W4383102502 hasConcept C128942645 @default.
- W4383102502 hasConcept C132525143 @default.
- W4383102502 hasConcept C152877465 @default.
- W4383102502 hasConcept C154945302 @default.
- W4383102502 hasConcept C162324750 @default.
- W4383102502 hasConcept C185592680 @default.
- W4383102502 hasConcept C2776214188 @default.
- W4383102502 hasConcept C2777615720 @default.
- W4383102502 hasConcept C41008148 @default.
- W4383102502 hasConcept C501734568 @default.
- W4383102502 hasConcept C539667460 @default.
- W4383102502 hasConcept C55493867 @default.
- W4383102502 hasConcept C80444323 @default.
- W4383102502 hasConceptScore W4383102502C104317684 @default.
- W4383102502 hasConceptScore W4383102502C119857082 @default.
- W4383102502 hasConceptScore W4383102502C124101348 @default.
- W4383102502 hasConceptScore W4383102502C128942645 @default.
- W4383102502 hasConceptScore W4383102502C132525143 @default.
- W4383102502 hasConceptScore W4383102502C152877465 @default.
- W4383102502 hasConceptScore W4383102502C154945302 @default.