Unsupervised learning of dislocation motion

Darren C. Pagan, Thien Q. Phan, Jordan S. Weaver, Austin R. Benson, Armand J. Beaudoin

Research output: Contribution to journalArticlepeer-review


The unsupervised learning technique, locally linear embedding (LLE), is applied to the analysis of X-ray diffraction data measured in-situ during the uniaxial plastic deformation of an additively manufactured nickel-based superalloy. With the aid of a physics-based material model, we find that the lower-dimensional coordinates determined using LLE appear to be physically significant and reflect the evolution of the defect densities that dictate strength and plastic flow behavior in the alloy. The implications of the findings for future constitutive model development are discussed, with a focus on wider applicability to microstructure evolution and phase transformation studies during in-situ materials processing.

Original languageEnglish (US)
Pages (from-to)510-518
Number of pages9
JournalActa Materialia
StatePublished - Dec 2019


  • Additive manufacturing
  • Machine learning
  • Nickel-based superalloy
  • Plasticity
  • X-ray diffraction

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Ceramics and Composites
  • Polymers and Plastics
  • Metals and Alloys


Dive into the research topics of 'Unsupervised learning of dislocation motion'. Together they form a unique fingerprint.

Cite this