Learning graph-matching edit-costs based on the optimality of the oracle's node correspondences

Cortés, Xavier; Serratosa, Francesc

doi:10.1016/j.patrec.2015.01.009

Cited by 48 publications

(24 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Two versions were utilized in the tests. R-1NN (Riesen and Bunke, 2009) and C-1NN (Cortés and Serratosa, 2015) where the values of K e = β e del = β e ins and K n = β n del = β n ins were borrowed from (Riesen and Bunke, 2010) and (Moreno-García et al, 2016), respectively. K n corresponds to the cost of a node deletion or insertion, and K e corresponds to the cost of an edge deletion or insertion.…”

Section: Experimental Settingmentioning

confidence: 99%

Learning error-correcting graph matching with a multiclass neural network

Martineau

Raveaux

Conte

et al. 2020

Pattern Recognition Letters

View full text Add to dashboard Cite

Section: Experimental Settingmentioning

confidence: 99%

Learning error-correcting graph matching with a multiclass neural network

Martineau

Raveaux

Conte

et al. 2020

Pattern Recognition Letters

View full text Add to dashboard Cite

“…We can separate these approaches in three main groups depending on its goal. By one hand, the first group (Raveaux R., 2017;Bunke, 2005, 2007;Leordeanu et al, 2012) focuses on the graph classification ratio, while the second group (Caetano et al, 2009;Cortés and Serratosa, 2016;Serratosa F., 2011;Cortés and Serratosa, 2015; ?) targets the node-to-node matching accuracy.…”

Section: State Of the Artmentioning

confidence: 99%

Learning edit cost estimation models for graph edit distance

Cortés

Conte

Cardot

2019

Pattern Recognition Letters

Self Cite

View full text Add to dashboard Cite

One of the most popular distance measures between a pair of graphs is the Graph Edit Distance. This approach consists of finding a set of edit operations that completely transforms a graph into another. Edit costs are introduced in order to penalize the distortion that each edit operation introduces. Then, one basic requirement when we design a Graph Edit Distance algorithm, is to define the appropriate edit cost functions. On the other hand, Machine Learning algorithms has been applied in many contexts showing impressive results, due, among other things, its ability to find correlations between input and output values. The aim of this paper is to bring the potentialities of Machine Learning to the Graph Edit Distance problem presenting a general framework in which this kind of algorithms are used to estimate the edit costs values for node substitutions.

show abstract

“…Several methods have been presented to learn the edit costs based on supervised machine learning techniques, which can be divided in two main groups. The ones that return a constant on the edit operations [20][21][22][23][24], and the other ones that define the edit costs as functions. For instance, in [25], they use a probabilistic model of the distribution of graph edit operations.…”

Section: Learning the Edit Costs And Graph Embeddingmentioning

confidence: 99%

Sub-optimal Graph Matching by Node-to-Node Assignment Classification

Cortés

Conte

Serratosa

2019

Graph-Based Representations in Pattern Recognition

Self Cite

View full text Add to dashboard Cite

In the recent years, Graph Edit Distance has awaken interest in the scientific community and some new graph-matching algorithms that compute it have been presented. Nevertheless, these algorithms usually cannot be used in real applications due to runtime restrictions. For this reason, other graph-matching algorithms have also been used that compute an approximation of the graph correspondence with lower runtime. Clearly, in a real application, there is a tradeoff between runtime and accuracy. One of the most costly part in these algorithms is the deduction of the node-to-node mapping. We present a new graph-matching algorithm that returns a graph correspondence without the explicit computation of the assignment problem. This is done thanks to a classification of the node-to-node assignment learned in a previous training stage.

show abstract

Learning graph-matching edit-costs based on the optimality of the oracle's node correspondences

Cited by 48 publications

References 35 publications

Learning error-correcting graph matching with a multiclass neural network

Learning error-correcting graph matching with a multiclass neural network

Learning edit cost estimation models for graph edit distance

Sub-optimal Graph Matching by Node-to-Node Assignment Classification

Contact Info

Product

Resources

About