Scalability and interpretability of graph neural networks for small molecules
This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections.
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Thesis |
| Language: | eng |
| Published: |
Massachusetts Institute of Technology
2019
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/1721.1/121639 |
| _version_ | 1826192996937236480 |
|---|---|
| author | Sangha, Manjot. |
| author2 | Joseph Jacobson. |
| author_facet | Joseph Jacobson. Sangha, Manjot. |
| author_sort | Sangha, Manjot. |
| collection | MIT |
| description | This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. |
| first_indexed | 2024-09-23T09:32:23Z |
| format | Thesis |
| id | mit-1721.1/121639 |
| institution | Massachusetts Institute of Technology |
| language | eng |
| last_indexed | 2024-09-23T09:32:23Z |
| publishDate | 2019 |
| publisher | Massachusetts Institute of Technology |
| record_format | dspace |
| spelling | mit-1721.1/1216392019-08-07T03:03:54Z Scalability and interpretability of graph neural networks for small molecules Sangha, Manjot. Joseph Jacobson. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science. Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science Electrical Engineering and Computer Science. This electronic version was submitted by the student author. The certified thesis is available in the Institute Archives and Special Collections. Thesis: M. Eng., Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science, 2018 Cataloged from student-submitted PDF version of thesis. Includes bibliographical references (pages 95-98). In this thesis I examine the use of graph neural networks for prediction tasks in chemistry with an emphasis on interpretable and scalable methods. I propose a novel kernel-inspired graph neural network architecture, called a subgraph matching neural network (SMNN), which is designed to have all feature representations and weights be human interpretable. I show that this network can achieve competitive performance with common graph neural network baselines. I also show that the network is capable of learning features that allow for transfer learning to larger molecules with significantly better performance than some baselines. This provides evidence the network is learning chemically useful representations. I then propose a framework for defining graph pooling operations to improve the scalability of graph neural networks with molecule size. I empirically examine some examples of these graph pooling layers and show that they can provide a significant speed-up without hurting accuracy, and even improving accuracy in some cases. Finally an instance of the SMNN network with a pooling layer is shown to achieve state-of-the-art accuracy on the Harvard Clean Energy Project dataset. by Manjot Sangha. M. Eng. M.Eng. Massachusetts Institute of Technology, Department of Electrical Engineering and Computer Science 2019-07-15T20:30:06Z 2019-07-15T20:30:06Z 2018 2018 Thesis https://hdl.handle.net/1721.1/121639 1098180075 eng MIT theses are protected by copyright. They may be viewed, downloaded, or printed from this source but further reproduction or distribution in any format is prohibited without written permission. http://dspace.mit.edu/handle/1721.1/7582 98 pages application/pdf Massachusetts Institute of Technology |
| spellingShingle | Electrical Engineering and Computer Science. Sangha, Manjot. Scalability and interpretability of graph neural networks for small molecules |
| title | Scalability and interpretability of graph neural networks for small molecules |
| title_full | Scalability and interpretability of graph neural networks for small molecules |
| title_fullStr | Scalability and interpretability of graph neural networks for small molecules |
| title_full_unstemmed | Scalability and interpretability of graph neural networks for small molecules |
| title_short | Scalability and interpretability of graph neural networks for small molecules |
| title_sort | scalability and interpretability of graph neural networks for small molecules |
| topic | Electrical Engineering and Computer Science. |
| url | https://hdl.handle.net/1721.1/121639 |
| work_keys_str_mv | AT sanghamanjot scalabilityandinterpretabilityofgraphneuralnetworksforsmallmolecules |