Benchmarking machine learning approaches for polarization mapping in ferroelectrics using 4D-STEM
Matej Martinc, Goran Dražić, Anton Kokalj, Katarina Žiberna, Janina Roknić, Matic Poberžnik, Sašo Džeroski, Andreja Benčan Golob · 2026
Read Benchmarking machine learning approaches for polarization mapping in ferroelectrics using 4D-STEM by Matej Martinc, Goran Dražić, Anton Kokalj, Katarina Žiberna, Janina Roknić, Matic Poberžnik, Sašo Džeroski, Andreja Benčan Golob (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/PMC13534660.txt.
Description
Matej Martinc and colleagues benchmark in 2026 machine learning models that map electric polarisation in ferroelectric materials from advanced electron microscope data. Such maps reveal function, yet reading them from four-dimensional diffraction data is hard by hand. Four model types were trained and compared on a niobate material.
Models trained on synthetic patterns scored high on matching synthetic tests, yet the gap to real experiments stayed a major barrier. A prototype training scheme plus classic dimension methods with augmentation looked most promising for closing that gap. Error patterns also matched crystal defects, hinting at defect detection uses.
Catalog record
| CREATOR | Matej Martinc, Goran Dražić, Anton Kokalj, Katarina Žiberna, Janina Roknić, Matic Poberžnik, Sašo Džeroski, Andreja Benčan Golob |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 121019 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |