A Metric Learning Approach to Graph Edit Costs for Regression

Jia, Linlin; Gaüzère, Benoît; Yger, Florian; Honeiné, Paul

doi:10.1007/978-3-030-73973-7_23

“…Since computing GED is an NP-hard problem [58], approximation methods have been proposed to estimate it. In this paper, we use a milestone method, called Bipartite [59], where the edit costs are optimized using a method from [60]. This approach boils down the quadratic assignment problem (QAP) to a linear one by finding a suboptimal edit path only based on local structural information.…”

Section: Feature Representation Statistics (S) Min Max Mean Stdmentioning

confidence: 99%

Predicting Redox Potentials by Graph-Based Machine Learning Methods

Jia,

Brémond,

Zaida

et al. 2024

Preprint

0

View full text Add to dashboard Cite

The evaluation of oxidation and reduction potentials is a pivotal task in various chemical fields. However, their accurate prediction by theoretical computations, which is a complementary task and sometimes the only alternative to experimental measurement, may be often resource-intensive and time-consuming. This paper addresses this challenge through the application of machine learning techniques, with a particular focus on graph-based methods (such as graph edit distances, graph kernels, and graph neural networks) that are reviewed to enlighten their deep links with theoreti- cal chemistry. To this aim, we establish the ORedOx159 database, a comprehensive, homogeneous (with reference values stemming from density functional theory calculations), and reliable resource containing 318 one-electron reduction and oxidation reactions and featuring 159 large organic com- pounds. Subsequently, we provide an instructive overview of the good practice in machine learning and of commonly utilized machine learning models. We then assess their predictive performances on the ORedOx159 dataset through extensive analyses. Our simulations using descriptors that are computed in an almost instantaneous way result in a notable improvement in prediction accuracy, with mean absolute error (MAE) values equal to 5.6 kcal mol−1 for reduction and 7.2 kcal mol−1 for oxidation potentials, which paves a way toward efficient in silico design of new electrochemical systems.

show abstract

“…To address this shortcoming, multiple supervised strategies are proposed. In [25], the edit costs are optimized by minimizing the difference between the GED and the distances between the prediction targets. While in [26], genetic algorithms are applied to optimize the costs for classification tasks.…”

Section: Related Workmentioning

confidence: 99%

“…Given T E the distances between prediction targets or ground truths, many edit costs optimization strategies fall into this framework, including the ones introduced in Section 2 (e.g., [26]). In the next section, we take advantage of a supervised metric learning strategy in [25] and define an instance of the framework.…”

Section: The Gecl Frameworkmentioning

confidence: 99%

Bridging Distinct Spaces in Graph-Based Machine Learning

Jia,

Ning,

Gaüzère

et al. 2023

Lecture Notes in Computer Science

Self Cite

1

0

View full text Add to dashboard Cite

Graph-based machine learning, encompassing Graph Edit Distances (GEDs), Graph Kernels, and Graph Neural Networks (GNNs), offers extensive capabilities and exciting potential. While each model possesses unique strengths for graph challenges, interrelations between their underlying spaces remain under-explored. In this paper, we introduce a novel framework for bridging these distinct spaces via GED cost learning. A supervised metric learning approach serves as an instance of this framework, enabling space alignment through pairwise distances and the optimization of edit costs. Experiments reveal the framework's potential for enhancing varied tasks, including regression, classification, and graph generation, heralding new possibilities in these fields.

show abstract

“…Since computing GED is an NP-hard problem [41], approximation methods have been proposed to estimate it. In this paper, we use a milestone method, Bipartite [42], where the edit costs are optimized using a method from [43]. This approach boils down the quadratic assignment problem (QAP) to a linear one by finding a suboptimal edit path only based on local structural information.…”

Section: Feature Representation Statistics (S) Min Max Mean Stdmentioning

confidence: 99%

Predicting Redox Potentials by Graph-Based Machine Learning Methods

Jia,

Brémond,

Zaida

et al. 2023

Preprint

0

View full text Add to dashboard Cite

The evaluation of oxidation and reduction potentials is a pivotal task in various chemical fields. However, their accurate prediction by theoretical computations, which is a complementary task and sometimes the only alternative to experimental measurement, may be often resource-intensive and time-consuming. This paper addresses this challenge through the application of machine learning techniques, with a particular focus on graph-based methods (such as graph edit distances, graph kernels, and graph neural networks) that are reviewed to enlighten their deep links with theoreti- cal chemistry. To this aim, we establish the ORedOx159 database, a comprehensive, homogeneous (with reference values stemming from density functional theory calculations), and reliable resource containing 318 one-electron reduction and oxidation reactions and featuring 159 large organic com- pounds. Subsequently, we provide an instructive overview of the good practice in machine learning and of commonly utilized machine learning models. We then assess their predictive performances on the ORedOx159 dataset through extensive analyses. Our simulations using descriptors that are computed in an almost instantaneous way result in a notable improvement in prediction accuracy, with mean absolute error (MAE) values equal to 5.6 kcal mol−1 for reduction and 7.2 kcal mol−1 for oxidation potentials, which paves a way toward efficient in silico design of new electrochemical systems.

show abstract

A Metric Learning Approach to Graph Edit Costs for Regression

Cited by 5 publications

References 27 publications

Predicting Redox Potentials by Graph-Based Machine Learning Methods

Predicting Redox Potentials by Graph-Based Machine Learning Methods

Bridging Distinct Spaces in Graph-Based Machine Learning

Predicting Redox Potentials by Graph-Based Machine Learning Methods

Contact Info

Product

Resources

About