Collection of functions to test optimization algorithms
Published:
A collection of continous functions (coded in Python) with known minimum to test optimization algorithms.
Posts by Collection
projects
Stochastic optimization algorithms
Published:
Several stochastic optimization methods (coded in Python) for derivative-free optimization of functions of n-dimensions.
Optimization algorithms used in ML
Published:
A pedagogical implementation (as .ipynb) of optimization algorithms used to trained ML models.
Aspen HYSYS-Python
Published:
An Aspen HYSYS-Python connection using spreadsheets.
Aspen Plus-Python
Published:
An Aspen Plus-Python connection using object paths.
Graph Neural Networks for the prediction of infinite dilution activity coefficients
Published:
A GNN model that predicts isothermal infinite dilution activity coefficients.
Gibbs–Helmholtz graph neural network: capturing the temperature dependency of activity coefficients at infinite dilution
Published:
A hybrid GNN model that predicts infinite dilution activity coefficients at varying temperatures.
Machine Learning in Chemical Engineering JupyterBook
Published:
A JupyterBook for the hands-on part of a Machine Learning in Chemical Engineering introductory course.
publications
Understanding the dynamic behaviour of semicontinuous distillation
Published in Computer Aided Chemical Engineering, 2018
In this paper we study the effect of the initial state of a ternary semi-continous distillation system on its periodic orbit.
Recommended citation: Madabhushi, P.B., Medina, E.I.S. and Adams II, T.A., 2018. Understanding the dynamic behaviour of semicontinuous distillation. In Computer Aided Chemical Engineering (Vol. 43, pp. 845-850). Elsevier. https://doi.org/10.1016/B978-0-444-64235-6.50148-0
Impacts of antiscalants on the formation of calcium solids: Implication on scaling potential of desalination concentrate
Published in Environmental Science: Water Research & Technology, 2019
Here, we investigate the effects of three widely used antiscalants (NTMP, EDTMP and DTPMP) on the precipitation of calcium from solutions under chemical conditions relevant to brackish desalination brine.
Recommended citation: Jain, T., Sanchez, E., Owens-Bennett, E., Trussell, R., Walker, S. and Liu, H., 2019. Impacts of antiscalants on the formation of calcium solids: Implication on scaling potential of desalination concentrate. Environmental Science: Water Research & Technology, 5(7), pp.1285-1294. https://doi.org/10.1039/C9EW00351G
Hybrid semi‐parametric modeling in separation processes: a review
Published in Chemie Ingenieur Technik, 2020
Here, we review some of the weaknesses in the current knowledge in separations design, simulation, optimization, and operation, and presents many examples where data‐driven and hybrid models have been used to facilitate these tasks.
Recommended citation: McBride, K., Sanchez Medina, E.I. and Sundmacher, K., 2020. Hybrid semi‐parametric modeling in separation processes: a review. Chemie Ingenieur Technik, 92(7), pp.842-855. https://doi.org/10.1002/cite.202000025
Acyclic modular flowsheet optimization using multiple trust regions and Gaussian process regression
Published in Computer Aided Chemical Engineering, 2021
Here, we present an algorithm to optimize process flowsheets using Gaussian processes regression and trust regions.
Recommended citation: Medina, E.S., Vallejo, D.R., Chachuat, B., Sundmacher, K., Petsagkourakis, P. and del Rio-Chanona, E.A., 2021. Acyclic modular flowsheet optimization using multiple trust regions and Gaussian process regression. In Computer Aided Chemical Engineering (Vol. 50, pp. 1117-1123). Elsevier. https://doi.org/10.1016/B978-0-323-88506-5.50172-8
Prediction of bioconcentration factors (BCF) using graph neural networks
Published in Computer Aided Chemical Engineering, 2021
Here, we present a Graph Neural Network approach for the prediction of bioconcentration factors (BCF) of small organic molecules.
Recommended citation: Medina, E.S., Linke, S. and Sundmacher, K., 2021. Prediction of bioconcentration factors (BCF) using graph neural networks. In Computer Aided Chemical Engineering (Vol. 50, pp. 991-997). Elsevier. https://doi.org/10.1016/B978-0-323-88506-5.50153-4
RaWaNet: Enriching Graph Neural Network Input via Random Walks on Graphs
Published in arXiv, 2021
Here, we propose a method for enriching the input information of Graph Neural Networks using a random walk data processing of the graphs based on three selected lengths.
Recommended citation: Iravanizad, A., Medina, E.I.S. and Stoll, M., 2021. RaWaNet: Enriching Graph Neural Network Input via Random Walks on Graphs. arXiv preprint arXiv:2109.07555. https://doi.org/10.48550/arXiv.2109.07555
Graph neural networks for the prediction of infinite dilution activity coefficients
Published in Digital Discovery, 2022
Here, we develop the first Graph Neural Network framework for predicting infinite dilution activity coefficients using an isothermal dataset.
Recommended citation: Medina, E.I.S., Linke, S., Stoll, M. and Sundmacher, K., 2022. Graph neural networks for the prediction of infinite dilution activity coefficients. Digital Discovery, 1(3), pp.216-225. https://doi.org/10.1039/D1DD00037C
Gibbs–Helmholtz graph neural network: capturing the temperature dependency of activity coefficients at infinite dilution
Published in Digital Discovery, 2023
Here, we develop a hybrid model combining graph neural networks with a Gibbs-Helmholtz derived expression for predicting limiting activity coefficients at varying temperatures.
Recommended citation: Medina, E.I.S., Linke, S., Stoll, M. and Sundmacher, K., 2023. Gibbs–Helmholtz graph neural network: capturing the temperature dependency of activity coefficients at infinite dilution. Digital Discovery. https://doi.org/10.1039/D2DD00142J
talks
Una introduccion a la optimizacion en Machine Learning (Spanish)
Published:
A brieve introduction to optimization algorithms such as gradient descent, RMSProp and Adam relevant in machine learning. Hands-on exercises using Google Colabs. Around 30 participants. This was one of the 2-hours tutorial sessions of the summer school of RIIAA 3.0.
Procesos Gaussianos y optimizacion Bayesiana (Spanish)
Published:
A brieve introduction to Gaussian processes and Bayesian optimization. Hands-on exercises using Google Colabs. Around 30 participants. This was one of the 2-hours tutorial sessions of the summer school of RIIAA 3.0.
Prediction of bioconcentration factors (BCF) using graph neural networks
Published:
I present this poster in the ESCAPE-31 conference.
Acyclic modular flowsheet optimization using multiple trust regions and Gaussian process regression
Published:
I present this poster in the ESCAPE-31 conference.
Introduccion a los metodos de optimizacion en Machine Learning (Spanish)
Published:
A brieve introduction to optimization methods used in machine learning. Hands-on exercises using Google Colabs. Around 30 participants. This was one of the 2-hours tutorial sessions of the summer school of RIIAA 4.0.
Optimizacion en Machine learning (Spanish)
Published:
An introduction to constrained optimization methods. Hands-on exercises using Google Colabs. Around 30 participants. This was one of the 2-hours tutorial sessions of the summer school of RIIAA 4.0.
Procesos Gaussianos y optimizacion Bayesiana (Spanish)
Published:
An introduction to Gaussian processes and Bayesian optimization. Hands-on exercises using Google Colabs. Around 30 participants. This was one of the 2-hours tutorial sessions of the summer school of RIIAA 4.0.
Basic Principles of Machine Learning: Regression and Classification
Published:
This talk was part of the Machine Learning for Process and Systems Engineering summer school at MPI-Magdeburg.
Optimizacion numerica (Spanish)
Published:
I presented this 1-week long workshop on numerical optimization together with Prof. Antonio del Rio at the Institute of Renewable Energy, UNAM.
Multi-Objective Bayesian optimization of process flowsheets using trust regions and quality set metrics.
Published:
I present this talk at the Data-Driven and Hybrid Modeling for Decision Making II session at the 2021 AIChE Annual Meeting.
Predicting Activity Coefficients at Infinite Dilution Using Hybrid Graph Neural Networks
Published:
I present this poster at the AIChE Annual Meeting 2022.
Prediction of activity coefficients at infinite dilution using hybrid Graph Neural Networks
Published:
I present this poster at the 14th IMPRS ProEng workshop.
Hybrid Graph Neural Networks for the prediction of activity coefficients
Published:
I present our work on hybrid GNNs applied to the prediction of activity coefficients. This work was presented to the DECHEMA expert group “Model-based process development and optimization”.
An introductory course of machine learning for chemical engineering students: a prototype
Published:
I present our work on the JupyterBook Machine Learning in Chemical Engineering.
Predicting activity coefficients at infinite dilution of polymer solutions using Graph Neural Networks
Published:
I present our work on the the use of Graph Neural Networks for the prediction of activity coeffcients at infinite dilution of polymer solutions.
Solvent pre-selection for extractive distillation using Gibbs-Helmholtz Graph Neural Networks
Published:
I present our work on the the use of the Gibbs-Helmholtz Graph Neural Network for solvent pre-selection in extractive distillation applications.
Molecular and mixture property prediction with Graph Neural Networks - Applications to chemical engineering
Published:
I was invited by Prof. Zavala to give this talk to the Department of Chemical & Biological Engineering at the University of Wisconsin-Madison.
Predicción de propiedades moleculares y de mezclas con Graph Neural Networks - Aplicaciones en ingeniería química
Published:
I was invited to give this seminar talk to the chemical engineering graduate students of the Insititute of Technology of Celaya.
Hybrid graph neural networks for fluid phase equilibria modeling
Published:
This was my tandem talk as part of the IMPRS ProEng program
A Comparison of the UNIFAC Model vs. Graph Neural Network-based Models for the Prediction of Binary Vapor-Liquid Equilibria
Published:
This was a talk at the 33rd European Symposium on Applied Thermodynamics 2024.
Surrogate modeling of dynamical systems: deep learning or model-order reduction?
Published:
This poster was presented at the Machine Learning for Chemistry and Chemical Engineering (ML4CCE) workshop of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 2024).
teaching
Machine Learning in Chemical Engineering (tutorials)
Masters-level course, Otto von Guericke University, 2021
I served as a teaching assistant (TA) for this Masters-level (1 semester-long) course delivering tutorials on ML applications to chemical engineering systems (e.g., thermophysical property prediction, fault detection, hybrid semi-parametric modeling). Semesters in which I have been TA:
- Summer semester 2020
- Summer semester 2021
Process Systems Engineering (tutorials)
Masters-level course, Otto von Guericke University, 2022
I served as a teaching assistant (TA) for this Masters-level (1 semester-long) course delivering tutorials on mass and energy balance and numerical methods for solving ODEs. Semesters in which I have been TA:
- Winter semester 2019
- Winter semester 2020
- Winter semester 2021
- Winter semester 2022
Machine Learning in Chemical Engineering (lectures)
Masters-level course, Otto von Guericke University, 2024
I served as the main lecturer for this Masters-level (1 semester-long) course introducing chemical engineering students to the field of ML (e.g., linear and generalized linear models, ANNs, kernel methods, hybrid semi-parametric modeling, clustering and dimensionality reduction). As part of the development of this course I lead the development of the Machine Learning in Chemical Engineering JupyterBook. Semesters in which I have been the main lecturer:
- Summer semester 2022
- Summer semester 2024