Machine-learning Based Extraction of the Short-Range Part of the Interaction in Non-contact Atomic Force Microscopy
arXiv:2004.04882 · doi:10.1063/5.0007754
Abstract
A machine-learning method for extracting the short-range part of the probe-surface interaction from force spectroscopy curves is presented. Our machine-learning algorithm consists of two stages: the first stage determines a boundary that separates the region where the short-range interaction is dominantly acting on the probe, and a second stage that finds the parameters to fit the interaction over the long-range region. We successfully applied this method to force spectroscopy maps acquired over the Si(111)-(7x7) surface and found, as a result, a faint structure on the short-range interaction for one of the probes used in the experiments that would have probably been obviated using human-supervised fitting strategies.
References in corpus (8)
- Adam: A Method for Stochastic Optimization
- Neural Machine Translation by Jointly Learning to Align and Translate
- An overview of gradient descent optimization algorithms
- A General-Purpose Machine Learning Framework for Predicting Properties of Inorganic Materials
- Learning agile and dynamic motor skills for legged robots
- Machine Learning Unifies the Modelling of Materials and Molecules
- Adversarial Examples that Fool both Computer Vision and Time-Limited Humans
- Automated Structure Discovery in Atomic Force Microscopy