Join us in welcoming Prof. Nazanin Bassiri-Gharb as she presents the first of her talks at the School of Maths & Physics, entitled 'Machine Learning for discovery of meaningful & Physical contributors to piezoresponse force microscopy'.
- Date(s)
- September 26, 2023
- Location
- Bell Lecture Theatre, Main Physics Building
- Time
- 15:30 - 17:00
Of interest mainly to staff & students of the School, the talk will take place on Tuesday 26th September at the Bell Lecture Theatre, Main Physics.
Refreshments will be available outside the Bell Lecture Theatre from 15:30. The talk itself will begin at 16:00 so feel free to come early for some tea, biscuits, and a chat!
Please confirm your attendance via this MS form – we look forward to seeing you there.
-----------
Abstract: From high dielectric permittivity and piezoresponse to the pyroelectric and electrocaloric response, ferroelectrics possess a broad range of functionalities. Complex and correlated characterization approaches at multiple length scales have been developed to tackle some of the most challenging questions about the origin of the functional response in ferroic and multiferroic materials. However, data generated through such methods is also of large complexity, and only allows limited insight from direct inspection and traditional statistical analysis. The recent developments in applications of machine learning techniques to materials science offer a path forward, providing powerful new approaches to handle and analyze the information. Specifically, clustering and dimensional reduction techniques are often used as immediate, low computation cost approaches to identify superimposed physical and chemical contributors to functional behavior within multidimensional datasets.
Here, I will provide an overview of machine learning methods applied to characterization of ferroelectric materials through piezoresponse force microscopy (PFM), and particularly the resonant PFM approaches. I will emphasize data curation that apply physical and chemical constraints to machine learning, and methods to mask or enhance different contributions to the signal. The discussed approaches are used on a path to separation and possible quantification of various chemical and physical contributors to the PFM response, enabling future use of this characterization method towards discovery of ferroelectricity in novel materials.