Marcos Gonçalves Quiles
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Marcos Gonçalves Quiles has a total of 173 co-authors. Here are the top co-authors by number of shared publications:
- Juarez L. F. Da Silva
University of São Paulo Av. Trabalhador São-Carlense 400
19 shared publications - Liang Zhao
University of São Paulo
5 shared publications - Ronaldo Cristiano Prati
Federal University of ABC
5 shared publications - Fabricio Breve
Sao Paulo State University
4 shared publications Felipe V. Calderan
Institute of Science and Technology, Federal University of São Paulo, Av. Cesare Lattes, 1201, 12247-014 São José dos Campos, SP, Brazil
4 shared publications- Gabriel A. Pinheiro
Federal University of São Paulo
4 shared publications
Latest Publications
- Open Access
- 23 October 2026
- Published by Springer Nature in International Journal of Data Science and Analytics
- Vol. 22 (1) , 353
A multivariate causal chain detection methodology assessed on the ENSO modulation of the Northeastern Brazil precipitation
This study presents a novel data-driven methodology for extraction causal chains from high-dimensional climate datasets, with a particular focus on detecting and visualizing the pathways that link climate anomalies across geographically distant regions. The proposed framework is designed to be computationally efficient while maintaining high interpretability, making it suitable for large-scale analytical applications in climate science. To assess its effectiveness, the methodology is applied to the well-studied teleconnection between the El Niño—Southern Oscillation and precipitation variability in Northeastern Brazil—a region highly susceptible to climate extremes such as droughts. The approach integrates clustering method, causal inference technique, and optimized path selection algorithm to construct interpretable causal chain graphs. Results demonstrate the methodology’s strong capability to identify key causal relationships and reproduce known climate phenomena. By effectively managing complex and massive climate data, this methodology contributes to the advancement of data science tools for climate analytics and offers valuable insights for understanding global climate teleconnections.
- 23 October 2026
- Published by Springer Nature in International Journal of Data Science and Analytics
- Vol. 22 (1) , 353
Open AccessA multivariate causal chain detection methodology assessed on the ENSO modulation of the Northeastern Brazil precipitation
This study presents a novel data-driven methodology for extraction causal chains from high-dimensional climate datasets, with a particular focus on detecting and visualizing the pathways that link climate anomalies across geographically distant regions. The proposed framework is designed to be computationally efficient while maintaining high interpretability, making it suitable for large-scale analytical applications in climate science. To assess its effectiveness, the methodology is applied to the well-studied teleconnection between the El Niño—Southern Oscillation and precipitation variability in Northeastern Brazil—a region highly susceptible to climate extremes such as droughts. The approach integrates clustering method, causal inference technique, and optimized path selection algorithm to construct interpretable causal chain graphs. Results demonstrate the methodology’s strong capability to identify key causal relationships and reproduce known climate phenomena. By effectively managing complex and massive climate data, this methodology contributes to the advancement of data science tools for climate analytics and offers valuable insights for understanding global climate teleconnections.
- Open Access
- 28 July 2026
- Published by Elsevier in Journal of Energy Storage
- Vol. 179
Electrode microstructure governs rate-dependent capacity retention: A metrology-aware optimization framework for sodium-ion batteries
- 28 July 2026
- Published by Elsevier in Journal of Energy Storage
- Vol. 179
Open AccessElectrode microstructure governs rate-dependent capacity retention: A metrology-aware optimization framework for sodium-ion batteries
- 4 July 2026
- Published by Springer Nature
- p. 204-217
Less is More: XAI-Driven Feature Selection for Efficient Molecular Property Prediction
- 4 July 2026
- Published by Springer Nature
- p. 204-217
Less is More: XAI-Driven Feature Selection for Efficient Molecular Property Prediction
- 1 July 2026
- Published by Springer Nature
- p. 493-510
A Compact QSAR Model for BRAF Inhibitors Based on Gaussian Process Regression
- 1 July 2026
- Published by Springer Nature
- p. 493-510
A Compact QSAR Model for BRAF Inhibitors Based on Gaussian Process Regression
- Open Access
- 26 January 2026
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 11 (5) , 7948-7958
Cut-SOAP: A Machine Learning Descriptor for Rapid Screening of Molecular Adsorption Energetics
Adsorption energy is a fundamental property in catalysis and chemical reaction studies; however, conventional quantum chemistry methods, such as density functional theory, provide high accuracy but are often computationally expensive or even impractical for screening large data sets or complex chemical systems. In this work, we proposed a machine learning (ML) pipeline that efficiently predicts relative energy interactions for molecular adsorption near the minimum molecule–cluster distance, at a fraction of the computational cost of quantum chemistry-based methods. Our approach begins by transforming the Fritz–Haber Institute ab initio materials simulation (FHI-aims) output data into feature arrays through a modified version of the Smooth Overlap of Atomic Positions (SOAP) descriptor, which we call Cut-SOAP. This modification reduces the dimensionality of the features by more than 97% while preserving most of the inherent quality of the data. With this method, we construct a large adsorption data set of more than 430,000 entries using real-world data. Then, a deep neural network was trained on this data set, analyzing the influence of architectural and hyperparameter choices on both computational cost and predictive accuracy. The model achieved a mean absolute error below 0.1 eV in the standard test set. To rigorously assess its generalization for real-world applications, we evaluated it on a challenging out-of-distribution data set, where it maintained a robust mean absolute error below 1.0 eV. The trained model is capable of making thousands of predictions in seconds, demonstrating the effectiveness of the pipeline for rapid screening. These results highlight the benefits of ML-based approaches for material screening, which offer accessible, efficient, and accurate tools for predicting relative energy interactions. This capability is a crucial step toward the accelerated discovery and optimization of catalytic systems.
- 26 January 2026
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 11 (5) , 7948-7958
Open AccessCut-SOAP: A Machine Learning Descriptor for Rapid Screening of Molecular Adsorption Energetics
Adsorption energy is a fundamental property in catalysis and chemical reaction studies; however, conventional quantum chemistry methods, such as density functional theory, provide high accuracy but are often computationally expensive or even impractical for screening large data sets or complex chemical systems. In this work, we proposed a machine learning (ML) pipeline that efficiently predicts relative energy interactions for molecular adsorption near the minimum molecule–cluster distance, at a fraction of the computational cost of quantum chemistry-based methods. Our approach begins by transforming the Fritz–Haber Institute ab initio materials simulation (FHI-aims) output data into feature arrays through a modified version of the Smooth Overlap of Atomic Positions (SOAP) descriptor, which we call Cut-SOAP. This modification reduces the dimensionality of the features by more than 97% while preserving most of the inherent quality of the data. With this method, we construct a large adsorption data set of more than 430,000 entries using real-world data. Then, a deep neural network was trained on this data set, analyzing the influence of architectural and hyperparameter choices on both computational cost and predictive accuracy. The model achieved a mean absolute error below 0.1 eV in the standard test set. To rigorously assess its generalization for real-world applications, we evaluated it on a challenging out-of-distribution data set, where it maintained a robust mean absolute error below 1.0 eV. The trained model is capable of making thousands of predictions in seconds, demonstrating the effectiveness of the pipeline for rapid screening. These results highlight the benefits of ML-based approaches for material screening, which offer accessible, efficient, and accurate tools for predicting relative energy interactions. This capability is a crucial step toward the accelerated discovery and optimization of catalytic systems.
- Open Access
- 6 January 2026
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 11 (2) , 2962-2975
Optimizing Molecular Descriptors for Reliable Adsorption Energy Prediction on Transition Metal Nanoclusters
Efficient catalytic processes are crucial for converting pollutant molecules into valuable products. Transition-metal nanoclusters show promise as a result of their tunable properties, but identifying active catalysts requires costly studies of the adsorption energetics. Machine learning offers a faster alternative, predicting adsorption energies when trained on suitable descriptors and relatively large density functional theory (DFT) data sets. This study evaluates the predictive power and transferability of two structural descriptors, the Coulomb matrix and the many-body tensor representation, on a diverse nanocluster-adsorbate data set using the random forest regression algorithm. Both descriptors achieved a mean absolute error of 0.05 eV in test data, but performance dropped significantly on an external generated set with unprecedented examples. Adding a simple electronic feature, the number of unpaired electrons of the adsorbate improved generalizability, even though with higher mean absolute errors compared to the original data set, highlighting the dependence on the training data.
- 6 January 2026
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 11 (2) , 2962-2975
Open AccessOptimizing Molecular Descriptors for Reliable Adsorption Energy Prediction on Transition Metal Nanoclusters
Efficient catalytic processes are crucial for converting pollutant molecules into valuable products. Transition-metal nanoclusters show promise as a result of their tunable properties, but identifying active catalysts requires costly studies of the adsorption energetics. Machine learning offers a faster alternative, predicting adsorption energies when trained on suitable descriptors and relatively large density functional theory (DFT) data sets. This study evaluates the predictive power and transferability of two structural descriptors, the Coulomb matrix and the many-body tensor representation, on a diverse nanocluster-adsorbate data set using the random forest regression algorithm. Both descriptors achieved a mean absolute error of 0.05 eV in test data, but performance dropped significantly on an external generated set with unprecedented examples. Adding a simple electronic feature, the number of unpaired electrons of the adsorbate improved generalizability, even though with higher mean absolute errors compared to the original data set, highlighting the dependence on the training data.
- Open Access
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 105-110
Towards Label Shift Adaptation for Robust IoT Device Identification
- 13 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 105-110
Open AccessTowards Label Shift Adaptation for Robust IoT Device Identification
- 29 September 2025
- Published by Springer Nature
- p. 426-442
Mitigating Data Scarcity in Polymer Property Prediction via Multi-task Auxiliary Learning
- 29 September 2025
- Published by Springer Nature
- p. 426-442
Mitigating Data Scarcity in Polymer Property Prediction via Multi-task Auxiliary Learning
- Open Access
- 13 September 2025
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 10 (37) , 42746-42759
Exploring the Adsorption Properties of Small Molecules on CeZr-Based Nanoclusters
An atomistic understanding of the interactions between the molecules and surfaces is crucial to optimize the catalytic performance in heterogeneous systems. In this investigation, calculations based on density functional theory are used to characterize the adsorption of a set of catalytically motivated molecules (CO, CO2, CH4, NH3, H2O, SO2) on ceria (Ce15O30), zirconia (Zr15O30), and mixed ceria-zirconia (Ce8Zr7O30) nanoclusters, which were selected based on their relevance in a variety of catalytic reactions. To obtain an improved analysis of all optimized adsorbed structures, we developed an automated algorithm to characterize the adsorption modes, covering the orientation and site preferences, based on the combination of Coulomb matrix representations with k-means clustering, using Silhouette scores to define the number of representative structures. From our calculations and analysis, we found that the 6 closest substrate atoms to the adsorbed molecule provide an optimal representation for the characterization of orientation and site preference of the selected molecules. The adsorption modes of CO, CO2, CH4, NH3, H2O, and SO2 were grouped into distinct classes, showing consistent orientation patterns, such as parallel or inclined geometries relative to the substrate. In general, the adsorption process does not induce large deformations in the oxide nanoclusters. In the lowest energy structures, the specific interaction preferences of the molecules with the oxide clusters follow the pattern: CO and NH3 form bonds via lone pairs on the C and N atoms, respectively; CH4 assumes the umbrella configuration; and H2O and SO2 interact through their O atoms. In particular, the SO2 molecule undergoes large changes in the bond angle, indicating a possible deformation toward the SO3 molecule, especially in ceria nanoclusters. With the exception of SO2, all remaining molecules contribute electron density to the substrate upon adsorption, whereas SO2 functions as an electron acceptor (Lewis acid).
- 13 September 2025
- Published by American Chemical Society (ACS) in ACS Omega
- Vol. 10 (37) , 42746-42759
Open AccessExploring the Adsorption Properties of Small Molecules on CeZr-Based Nanoclusters
An atomistic understanding of the interactions between the molecules and surfaces is crucial to optimize the catalytic performance in heterogeneous systems. In this investigation, calculations based on density functional theory are used to characterize the adsorption of a set of catalytically motivated molecules (CO, CO2, CH4, NH3, H2O, SO2) on ceria (Ce15O30), zirconia (Zr15O30), and mixed ceria-zirconia (Ce8Zr7O30) nanoclusters, which were selected based on their relevance in a variety of catalytic reactions. To obtain an improved analysis of all optimized adsorbed structures, we developed an automated algorithm to characterize the adsorption modes, covering the orientation and site preferences, based on the combination of Coulomb matrix representations with k-means clustering, using Silhouette scores to define the number of representative structures. From our calculations and analysis, we found that the 6 closest substrate atoms to the adsorbed molecule provide an optimal representation for the characterization of orientation and site preference of the selected molecules. The adsorption modes of CO, CO2, CH4, NH3, H2O, and SO2 were grouped into distinct classes, showing consistent orientation patterns, such as parallel or inclined geometries relative to the substrate. In general, the adsorption process does not induce large deformations in the oxide nanoclusters. In the lowest energy structures, the specific interaction preferences of the molecules with the oxide clusters follow the pattern: CO and NH3 form bonds via lone pairs on the C and N atoms, respectively; CH4 assumes the umbrella configuration; and H2O and SO2 interact through their O atoms. In particular, the SO2 molecule undergoes large changes in the bond angle, indicating a possible deformation toward the SO3 molecule, especially in ceria nanoclusters. With the exception of SO2, all remaining molecules contribute electron density to the substrate upon adsorption, whereas SO2 functions as an electron acceptor (Lewis acid).
- Open Access
- 12 March 2025
- Published by Qeios Ltd
Review of: "Unified Guidance for Geometry-Conditioned Molecular Generation"
This paper introduces a framework (UniGuide) that applies self-guidance techniques to an unconditional molecular diffusion model to flexibly incorporate three-dimensional geometric constraints. Rather than training specialized, condition-specific mod...
- 12 March 2025
- Published by Qeios Ltd
Open AccessReview of: "Unified Guidance for Geometry-Conditioned Molecular Generation"
This paper introduces a framework (UniGuide) that applies self-guidance techniques to an unconditional molecular diffusion model to flexibly incorporate three-dimensional geometric constraints. Rather than training specialized, condition-specific mod...