Curiosity-driven sampling, which measures where a neural network predicts it will make the largest spectral errors, learns structure-property maps faster than random sampling in ferroelectric microscopy.
Spiral high-speed scanning tunneling microscopy: Tracking atomic diffusion on the millisecond timescale
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.mtrl-sci 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Curiosity Driven Exploration to Optimize Structure-Property Learning in Microscopy
Curiosity-driven sampling, which measures where a neural network predicts it will make the largest spectral errors, learns structure-property maps faster than random sampling in ferroelectric microscopy.