Learning to grow: Control of material self-assembly using evolutionary reinforcement learning

被引:39
作者
Whitelam, Stephen [1 ]
Tamblyn, Isaac [2 ,3 ]
机构
[1] Lawrence Berkeley Natl Lab, Mol Foundry, 1 Cyclotron Rd, Berkeley, CA 94720 USA
[2] Natl Res Council Canada, Ottawa, ON, Canada
[3] Vector Inst Artificial Intelligence, Toronto, ON, Canada
关键词
CONTROLLING CRYSTALLIZATION; PATCHY; DESIGN; MECHANISMS; PATHWAYS; GAME; GO;
D O I
10.1103/PhysRevE.101.052604
中图分类号
O35 [流体力学]; O53 [等离子体物理学];
学科分类号
070204 ; 080103 ; 080704 ;
摘要
We show that neural networks trained by evolutionary reinforcement learning can enact efficient molecular self-assembly protocols. Presented with molecular simulation trajectories, networks learn to change temperature and chemical potential in order to promote the assembly of desired structures or choose between competing polymorphs. In the first case, networks reproduce in a qualitative sense the results of previously known protocols, but faster and with higher fidelity; in the second case they identify strategies previously unknown, from which we can extract physical insight. Networks that take as input the elapsed time of the simulation or microscopic information from the system are both effective, the latter more so. The evolutionary scheme we have used is simple to implement and can be applied to a broad range of examples of experimental self-assembly, whether or not one can monitor the experiment as it proceeds. Our results have been achieved with no human input beyond the specification of which order parameter to promote, pointing the way to the design of synthesis protocols by artificial intelligence.
引用
收藏
页数:13
相关论文
共 71 条
[1]  
[Anonymous], 2014, Markov decision processes: discrete stochastic dynamic programming
[2]  
[Anonymous], 2013, P ADV NEUR INF PROC
[3]  
[Anonymous], 2016, PROC INT C MACH LEAR
[4]  
[Anonymous], 2017, ARXIV171206564
[5]   The Arcade Learning Environment: An Evaluation Platform for General Agents [J].
Bellemare, Marc G. ;
Naddaf, Yavar ;
Veness, Joel ;
Bowling, Michael .
JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH, 2013, 47 :253-279
[6]   Colloidal interactions and self-assembly using DNA hybridization [J].
Biancaniello, PL ;
Kim, AJ ;
Crocker, JC .
PHYSICAL REVIEW LETTERS, 2005, 94 (05)
[7]   Theoretical and numerical study of the phase diagram of patchy colloids: Ordered and disordered patch arrangements [J].
Bianchi, Emanuela ;
Tarlaglia, Piero ;
Zaccarelli, Emanuela ;
Sciortino, Francesco .
JOURNAL OF CHEMICAL PHYSICS, 2008, 128 (14)
[8]  
Brockman G., 2016, OPENAI GYM
[9]   Pharmaceutical Crystallization [J].
Chen, Jie ;
Sarma, Bipul ;
Evans, James M. B. ;
Myerson, Allan S. .
CRYSTAL GROWTH & DESIGN, 2011, 11 (04) :887-895
[10]  
Conti E., 2018, Advances in Neural Information Processing Systems