Generating High-Temperature Polymer Dielectrics with a Graph Convolution and Reinforcement Learning Framework
Source abstract
Abstract
Generating molecular structures with certain desired properties is a fundamental problem for material science research. Deep learning models have demonstrated big potential to identify novel materials by exploring large chemical spaces efficiently. Here, we demonstrate the generation of hypothetical polymers with a high glass-transition temperature ( T g ) and a wide band gap ( E g ) using a general graph convolutional network-based model through reinforcement learning. Two forward property prediction models on T g and E g are employed to provide domain-specific rewards. We further incorporate prior knowledge in the form of adversarial rewards given by a generative adversarial network framework via a data set of exemplar polymer structures. The domain-specific rewards and adversarial rewards are optimized in a reinforcement learning framework through policy gradient methods. Rules of chemistry are incorporated in the molecule generation environment as hard constraints. The generation process is guided to generate promising high-temperature polymer dielectrics for energy storage applications by property targeting and constrained property optimization. This work demonstrates potential for exploring large chemical spaces for the discovery of polymers with desirable properties.
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Dielectric materials and actuators · Advanced Sensor and Energy Harvesting Materials · Synthesis and properties of polymers
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
State Key Laboratory of Power System Operation and Control, Department of Electrical Engineering
Tsinghua University
Funding
National Key Research and Development Program of China · National Natural Science Foundation of China
References (39)
Show source-provided references
1. 10.1002/(SICI)1098-1128(199601)16:1<3::AID-MED1>3.0.CO;2-610.1002/(SICI)1098-1128(199601)16:1<3::AID-MED1>3.0.CO;2-6 ↗
2. 10.1186/s13321-017-0235-x10.1186/s13321-017-0235-x ↗
3. 10.1021/acscentsci.7b0051210.1021/acscentsci.7b00512 ↗
4. Kusner M. J. · 2017 · 1945
5. 10.1021/acscentsci.7b0057210.1021/acscentsci.7b00572 ↗
6. 10.1021/acs.chemmater.0c0333210.1021/acs.chemmater.0c03332 ↗
7. De Cao, N.; Kipf, T. MolGAN: An implicit generative model for small molecular graphs. arXiv 2018, arXiv.1805.11973.
8. You J. · 2018 · 6410
9. 10.1186/s13321-020-00454-310.1186/s13321-020-00454-3 ↗
10. 10.1021/ci00057a00510.1021/ci00057a005 ↗
11. 10.1021/acscentsci.7b0030310.1021/acscentsci.7b00303 ↗
12. 10.1021/acs.chemrev.1c0079310.1021/acs.chemrev.1c00793 ↗
13. 10.1021/acs.chemrev.2c0023110.1021/acs.chemrev.2c00231 ↗
14. 10.1016/j.progpolymsci.2019.06.00310.1016/j.progpolymsci.2019.06.003 ↗
15. 10.1146/annurev-matsci-070317-12443510.1146/annurev-matsci-070317-124435 ↗
16. 10.1039/D0CS00765J10.1039/D0CS00765J ↗
17. 10.1002/adma.20220362310.1002/adma.202203623 ↗
18. 10.1063/5.002375910.1063/5.0023759 ↗
19. 10.1016/j.patter.2021.10023810.1016/j.patter.2021.100238 ↗
20. 10.1021/acs.chemmater.2c0299110.1021/acs.chemmater.2c02991 ↗
21. 10.1038/s41524-019-0203-210.1038/s41524-019-0203-2 ↗
22. Otsuka S. · 2011 · 2210.1109/EIDWT.2011.13 ↗
23. 10.1021/acs.jcim.0c0072610.1021/acs.jcim.0c00726 ↗
24. 10.1063/5.009474610.1063/5.0094746 ↗
25. 10.1021/acs.jcim.3c0157210.1021/acs.jcim.3c01572 ↗
26. 10.1186/1758-2946-1-810.1186/1758-2946-1-8 ↗
27. Goodfellow I. · 2014 · 2672
28. Schulman, J.; Wolski, F.; Dhariwal, P.; Radford, A.; Klimov, O. Proximal policy optimization algorithms. arXiv 2017, arXiv.1707.06347.
29. Hamilton W. · 2017 · 1025
30. Schulman, J.; Moritz, P.; Levine, S.; Jordan, M.; Abbeel, P. High-dimensional continuous control using generalized advantage estimation. arXiv 2015, arXiv.1506.02438.
31. Levine S. · 2013
32. Brockman, G.; Cheung, V.; Pettersson, L.; Schneider, J.; Schulman, J.; Tang, J.; Zaremba, W. Openai gym. arXiv 2016, arXiv.1606.01540.
33. Landrum G. · GitHub and SourceForge · 2016 · 10 · 3592822
34. Ioffe S. · 2015 · 448
35. 10.1021/ci100050t10.1021/ci100050t ↗
36. Khazana. A computational materials knowledgebase https://khazana.gatech.edu/ (accessed Mar 20, 2024).
37. 10.1002/adma.20200049910.1002/adma.202000499 ↗
38. 10.1038/s41467-024-50413-x10.1038/s41467-024-50413-x ↗
39. 10.1038/s41560-024-01670-z10.1038/s41560-024-01670-z ↗