BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Machine learning clustering technique applied to X-ray diffraction
  patterns to distinguish alloy substitutions -  Ryo Maezono \, JAIST\, Ish
 ikawa\, Japan
DTSTART:20190911T103000Z
DTEND:20190911T113000Z
UID:TALK128788@talks.cam.ac.uk
CONTACT:77905
DESCRIPTION:SmFe12 is one of the candidate of the main phase in rare-earth
  permanent magnets [1]. The origin of intrinsic properties emerging at hig
 h temperature as well as that of the phase stability has not yet been clar
 ified well. Introducing Ti and Zr to substitute Fe and Sm is found to impr
 ove the magnetic properties and the phase stability. To clarify the mechan
 ism how the substitutions improve the properties\, it is desired to identi
 fy substituted sites and its amount quantitatively\, preferably with high 
 throughput efficiency for accelerating the ‘materials tuning’. Motivat
 ed by the above\, we have developed [2] a machine learning clustering tech
 nique to distinguish powder XRD patterns to get such microscopic identific
 ations about the atomic substitutions. Ab initio calculations are used to 
 generate supervising references for the machine learning of XRD patterns: 
 We prepared several possible model structures with substituents located on
  each different sites over a range of substitution fractions. Geometrical 
 optimizations for each model give slight different structures each other. 
 Then we generated many XRD patterns calculated from each structure. We fou
 nd that the DTW (dynamic time wrapping) analysis can capture slight shifts
  in XRD peak positions corresponding to the differences of each relaxed st
 ructure\, distinguishing the fractions and positions of substituents. We h
 ave established such a clustering technique using Ward’s analysis on top
  of the DTW \, being capable to sort out simulated XRD patterns based on t
 he distinction. The established technique can hence learn the corresponden
 ce between XRD peak shifts and microscopic structures with substitutions o
 ver many supervising simulated data. Since the ab initio simulation can al
 so give several properties such as magnetization for each structure\, the 
 correspondence in the machine learning can further predict functional prop
 erties of materials when it is applied to the experimental XRD patterns\, 
 not only being capable to distinguish the atomic substitutions. The ‘mac
 hine learning technique for XRD patterns’ developed here has therefore t
 he wider range of applications not limited only on magnets\, but further o
 n those materials which properties are tuned by the atomic substitutions.\
 n\nWe also provide our updated challenges using deep learning technique ap
 plied to XRD patterns.\n\nREFERENCES\n\n1. K. Kobayashi et al.\, J. Magn. 
 Magn. Mater. 426\, 273 (2017).\n\n2. K. Utimula\, R. Hunkao\, M. Yano\, H.
  Kimoto\, K. Hongo\, S. Kawaguchi\, S.Suwanna\, R. Maezono\, arXiv:1810.03
 972.
LOCATION:TCM Seminar Room\, Cavendish Laboratory
END:VEVENT
END:VCALENDAR