WiSe 23/24: Machine Learning for Materials Science
Philipp Florian Benner, Alexander Karl Kister
Comments
This seminar introduces data science students to the rapidly
evolving field of materials science. Materials science involves the
study of the properties, structure, and behavior of various
materials, ranging from metals and polymers to semiconductors and
biomaterials. The ultimate goal of materials science is to design and
develop new materials with the specific properties and functionalities
that are needed to meet the requirements of a given target application,
such as energy storage, electronics, medicine, and more.
In recent years, machine learning has emerged as a powerful tool for
materials science researchers to accelerate the discovery and design
of new materials. Machine learning methods are used to analyze large
datasets of materials properties, predict new materials with desired
properties, and optimize the synthesis and processing of
materials. This seminar will explore a diverse set of machine learning
methods and their applications in materials science.
The seminar will begin with an introduction to the goals and
challenges of materials science, including the importance of materials
in various applications and the need for faster and more efficient
materials discovery. Next, students will select a pair of papers from
the literature, one paper focuses on a materials science application
and the other on a deeper discussion of the used ML method.
Each student will present a summary of the two papers: The
presentations should be structured such that 40% of the time is
devoted to presenting the materials science application, 50% of the
time for discussing the machine learning method, and 10% of the time
for presenting the results of the study. After the talks there will be
a discussion, which is lead by a moderator. Each student will be the
moderator of one of the discussions.
By the end of this seminar, students will have a broad understanding
of the current state-of-the-art in machine learning for materials
science and will be able to critically evaluate and apply machine
learning methods to materials science datasets. This seminar is ideal
for data science students who are interested in materials science,
chemistry, physics, or engineering, and want to gain insights in
applying machine learning to real-world materials science problems.
close
16 Class schedule
Regular appointments