CRYSTAL: a multi-agent AI system for automated mapping of materials’ crystal structures

We introduce CRYSTAL, a multi-agent AI system for crystal-structure phase mapping. CRYSTAL is the first system that can automatically generate a portfolio of physically meaningful phase diagrams for expert-user exploration and selection. CRYSTAL outperforms previous methods to solve the example Pd-Rh-Ta phase diagram, enabling the discovery of a mixed-intermetallic methanol oxidation electrocatalyst. The integration of multiple data-knowledge sources and learning and reasoning algorithms, combined with the exploitation of problem decompositions, relaxations, and parallelism, empowers AI to supersede human scientific data interpretation capabilities and enable otherwise inaccessible scientific discovery in materials science and beyond.

[1]  Reddington,et al.  Combinatorial electrochemistry: A highly parallel, optical screening method for discovery of better electrocatalysts , 1998, Science.

[2]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[3]  H. Abruña,et al.  Electrocatalytic oxidation of formic acid at an ordered intermetallic PtBi surface. , 2003, Chemphyschem : a European journal of chemical physics and physical chemistry.

[4]  Christopher H. Bryant,et al.  Functional genomic hypothesis generation and experimentation by a robot scientist , 2004, Nature.

[5]  Paris Smaragdis,et al.  Non-negative Matrix Factor Deconvolution; Extraction of Multiple Sound Sources from Monophonic Inputs , 2004, ICA.

[6]  Lei Zhang,et al.  A review of anode catalysis in the direct methanol fuel cell , 2006 .

[7]  R. V. Dover,et al.  A high-throughput search for direct methanol fuel cell anode electrocatalysts of type PtxBiyPbz , 2007 .

[8]  Manuel Moliner,et al.  A reliable methodology for high throughput identification of a mixture of crystallographic phases from powder X-ray diffraction data , 2008 .

[9]  Claudio Bianchini,et al.  Palladium-based electrocatalysts for alcohol oxidation in half cells and in direct alcohol fuel cells. , 2009, Chemical reviews.

[10]  Alexander Kazimirov,et al.  High energy x-ray diffraction/x-ray fluorescence spectroscopy for high-throughput analysis of composition spread thin films. , 2009, The Review of scientific instruments.

[11]  I Takeuchi,et al.  Rapid identification of structural phases in combinatorial thin-film libraries using x-ray diffraction and non-negative matrix factorization. , 2009, The Review of scientific instruments.

[12]  Andrzej Cichocki,et al.  Nonnegative Matrix and Tensor Factorization T , 2007 .

[13]  John M. Gregoire,et al.  Improved Fuel Cell Oxidation Catalysis in Pt1−xTax† , 2010 .

[14]  Ronan Le Bras,et al.  Constraint Reasoning and Kernel Clustering for Pattern Decomposition with Scaling , 2011, CP.

[15]  R. V. Dover,et al.  High Throughput Thin Film Pt-M Alloys for Fuel Electrooxidation: Low Concentrations of M (M = Sn, Ta, W, Mo, Ru, Fe, In, Pd, Hf, Zn, Zr, Nb, Sc, Ni, Ti, V, Cr, Rh) , 2012 .

[16]  Manh Cuong Nguyen,et al.  On-the-fly machine-learning for high-throughput experiments: search for rare-earth-free permanent magnets , 2014, Scientific Reports.

[17]  Jelena Stajic,et al.  Artificial intelligence. Rise of the Machines. , 2015, Science.

[18]  I Takeuchi,et al.  High-throughput determination of structural phase diagram and constituent phases using GRENDEL , 2015, Nanotechnology.

[19]  Rahul Rao,et al.  Autonomy in materials research: a case study in carbon nanotube growth , 2016 .

[20]  John M. Gregoire,et al.  Perspective: Composition–structure–property mapping in high-throughput experiments: Turning data into knowledge , 2016 .

[21]  Jennifer N. Wei,et al.  Use machine learning to find energy materials. , 2017 .

[22]  Chiho Kim,et al.  Machine learning in materials informatics: recent applications and prospects , 2017, npj Computational Materials.

[23]  Alán Aspuru-Guzik,et al.  Use machine learning to find energy materials  , 2017, Nature.

[24]  Santosh K. Suram,et al.  Relaxation Methods for Constrained Matrix Factorization Problems: Solving the Phase Mapping Problem in Materials Discovery , 2017, CPAIOR.

[25]  Ronan Le Bras,et al.  Automated Phase Mapping with AgileFD and its Application to Light Absorber Discovery in the V-Mn-Nb Oxide System. , 2016, ACS combinatorial science.

[26]  Eric Smalley,et al.  AI-powered drug discovery captures pharma interest , 2017, Nature Biotechnology.

[27]  Andrew J. Medford,et al.  Scaling-Relation-Based Analysis of Bifunctional Catalysis: The Case for Homogeneous Bimetallic Alloys , 2017 .

[28]  Demis Hassabis,et al.  Mastering the game of Go without human knowledge , 2017, Nature.

[29]  Ichiro Takeuchi,et al.  Fulfilling the promise of the materials genome initiative with high-throughput experimental methodologies , 2017 .

[30]  Alán Aspuru-Guzik,et al.  Accelerating the discovery of materials for clean energy in the era of smart automation , 2018, Nature Reviews Materials.

[31]  Ichiro Takeuchi,et al.  Unsupervised phase mapping of X-ray diffraction data by nonnegative matrix factorization integrated with custom clustering , 2018, npj Computational Materials.