Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Concept Videos

Reinforcements in Concrete01:25

Reinforcements in Concrete

367
Reinforced concrete is a composite material used extensively in construction, combining the compressive strength of concrete with the tensile strength of steel. This synergy is essential as concrete, while excellent at resisting compression, is weak under tension. Steel bars, or rebars, are embedded in the concrete to handle these tensile forces. The choice of steel is strategic; it shares a similar coefficient of thermal expansion with concrete, which ensures uniformity in response to...
367
Reinforcement01:23

Reinforcement

722
Positive and negative reinforcement are key concepts in operant conditioning, a learning process where the consequences of a behavior affect the likelihood of that behavior being repeated.
Positive reinforcement occurs when a behavior is followed by the presentation of a rewarding stimulus, increasing the frequency of that behavior. For example:
722

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Active Δ-learning with universal potentials for global structure optimization.

Physical chemistry chemical physics : PCCP·2025
Same author

Anharmonic infrared spectra of cationic pyrene and superhydrogenated derivatives.

The Journal of chemical physics·2025
Same author

Δ-model correction of foundation model based on the model's own understanding.

The Journal of chemical physics·2025
Same author

Accelerating structure search using atomistic graph-based classifiers.

The Journal of chemical physics·2024
Same author

Cascading symmetry constraint during machine learning-enabled structural search for sulfur-induced Cu(111)-(43×43) surface reconstruction.

The Journal of chemical physics·2024
Same author

Structure of an Ultrathin Oxide on Pt<sub>3</sub>Sn(111) Solved by Machine Learning Enhanced Global Optimization.

Angewandte Chemie (Weinheim an der Bergstrasse, Germany)·2024

Related Experiment Video

Updated: Dec 21, 2025

Photorealistic Learned Landscapes for Augmented Reality
06:54

Photorealistic Learned Landscapes for Augmented Reality

Published on: June 27, 2025

587

Structure prediction of surface reconstructions by deep reinforcement learning.

Søren A Meldgaard1, Henrik L Mortensen1, Mathias S Jørgensen1

  • 1Department of Physics and Astronomy, Aarhus University, DK-8000 Aarhus C, Denmark.

Journal of Physics. Condensed Matter : an Institute of Physics Journal
|May 21, 2020
PubMed
Summary

This study combines image recognition and reinforcement learning to determine atomistic surface structures. The AI agent learns to build optimal surface reconstructions for titanium and tin oxides using quantum mechanical calculations.

Keywords:
chemical physicsglobal optimizationimage recognitionreinforcement learningstructure predictionsurface reconstruction

More Related Videos

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

926

Related Experiment Videos

Last Updated: Dec 21, 2025

Photorealistic Learned Landscapes for Augmented Reality
06:54

Photorealistic Learned Landscapes for Augmented Reality

Published on: June 27, 2025

587
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
03:31

Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications

Published on: December 15, 2023

926

Area of Science:

  • Materials Science
  • Computational Chemistry
  • Artificial Intelligence

Background:

  • Determining atomistic surface structures is crucial for understanding material properties.
  • Traditional methods can be computationally expensive and time-consuming.
  • Surface reconstruction significantly impacts material performance.

Purpose of the Study:

  • To develop an AI-driven approach for determining atomistic structures of reconstructed crystalline surfaces.
  • To leverage image recognition and reinforcement learning for efficient surface structure prediction.
  • To investigate the feasibility of this method for complex oxide surfaces.

Main Methods:

  • A deep neural network was employed as a reinforcement learning agent.
  • The agent interacted with an environment featuring a density functional theory program for quantum mechanical evaluations.
  • The 3D atomistic structure was processed as stacked 2D images, with the agent deciding atom placement and site occupancy.

Main Results:

  • The reinforcement learning agent successfully learned to build optimal surface reconstructions for anatase TiO2(001)-(1 × 4) and rutile SnO2(110)-(4 × 1).
  • The process required between 1,000 and 10,000 single-point density functional theory evaluations per surface.
  • This demonstrates the potential of AI in predicting complex surface reconstructions.

Conclusions:

  • Image recognition and reinforcement learning offer a powerful combination for predicting atomistic surface structures.
  • The developed AI agent can autonomously learn to construct optimal surface reconstructions.
  • This approach shows promise for accelerating materials discovery and design.