Email RIS file
Bookmark
Report an Issue

Manifold Learning in Structural Biology : Docking, Folding, and Simulation

Author
Sha, Congzhou
Published
[University Park, Pennsylvania] : Pennsylvania State University, 2024.
Physical Description
1 electronic document
Additional Creators
Dokholyan, Nikolay V.
Access Online

Availability

I Want It
I Want It

Finding items...

Graduate Program
Restrictions on Access
Open Access.
Summary
Machine learning and neural networks are used in many areas of science to perform data fitting, automation, and inference. However, these methods are often used as black boxes, leading to uncertainty on the scientist's part in their accuracy and robustness. Many problems in structural biology may be addressed using machine learning techniques, from virtual drug docking and RNA 3D structure to molecular dynamics. In order to use machine learning effectively, I have constructed principled and rational methods based on physics, geometry, and symmetry to address problems in structural biology. By taking advantage of fundamental mathematical properties in the problems under consideration, we are able to (1) develop efficient, accurate, and trustworthy inferential models and (2) abstract away from the details of the problem and develop general methods applicable to other areas of science. The choices I have made in constructing these methods reflect our current understanding of deep learning, the structure of scientific data, and the mathematical principles underlying our models of Nature.
Other Subject(s)
Genre(s)
Dissertation Note
Ph.D. Pennsylvania State University 2024.
Technical Details
The full text of the dissertation is available as an Adobe Acrobat .pdf file ; Adobe Acrobat Reader required to view the file.
View MARC record | catkey: 44541231