Project Info
MACHINE LEARNING-DRIVEN DISCOVERY OF ADSORBENTS FOR CO2 CAPTURE
Diego Gomez-Gualdron
dgomezgualdron@mines.edu
Project Goals and Description:
The current project will look to develop machine learning models that can predict the CO2 capture properties of millions of candidate compositions of nanoporous materials, whose "prototypes" have been built computationally by tools previously developed in my group. Among these materials we expect to find some materials with record-breaking CO2 capture properties, but first we need to develop the tools to be able to predict "instantaneously" (like machine learning does) what these properties for each prototype are.
From the application point of view, the goal is interesting because it could have an impact on development of CO2 capture technologies. From the methodological point of view, the goal is interesting because we will look for ways to solve the so-called "data scarcity" issue as we attempt to develop the machine learning models that will make the CO2 capture prediction.
More Information:
Grand Challenge: Develop carbon sequestration methods.
To learn about computational methods applied to discovery of materials for CO2 capture:
https://www.science.org/doi/full/10.1126/sciadv.1600909
https://pubs.rsc.org/en/content/articlelanding/2012/ee/c2ee23201d
To learn about CO2 capture in materials via machine learning/data science:
https://pubs.acs.org/doi/abs/10.1021/acs.chemmater.8b02257
https://www.nature.com/articles/s41586-019-1798-7
Primary Contacts:
dgomezgualdron@mines.edu
Student Preparation
Qualifications
-Desire/willingness/curiosity to work with computers and gain coding skills
-Desire/willingness/curiosity to learn and do machine learning work
-A good foundation from calculus, chemistry and physics courses (and also perhaps thermodynamics).
TIME COMMITMENT (HRS/WK)
5 hours per week
SKILLS/TECHNIQUES GAINED
The student will gain the following skills:
-Working with UNIX operating systems
-Writing short codes in Bash and Python
-Learn to train machine learning models in Python
-Handle and visualize large datasets in Python
-Understand the basics of adsorption and CO2 sequestration technologies
-Understanding the basics of molecular simulation
MENTORING PLAN
-Weekly meeting with Prof. Gomez-Gualdron
-As needed, regular training by grad students.
Preferred Student Status
Sophomore
Junior