Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

Associative Learning

Associative Learning

Associative learning is a fundamental concept in behavioral psychology, wherein a connection is established between two stimuli or events, leading to a learned response. This process is critical in understanding how behaviors are acquired and modified. Conditioning, the mechanism through which associations are formed, can be divided into two main types: classical conditioning and operant conditioning, each elucidating different aspects of associative learning.
Classical conditioning, also known...

You might also read

Related Articles

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

Sort by

Same author

Programmable Surface Catalyzed Heterogeneous Nucleation Enables "Double-Cable" Light-Harvesting Supramolecular Polymers.

Journal of the American Chemical Society·2026

Same author

Mechanistic insights into 50<i>S</i> precursor recognition and targeting by erythromycin resistance methyltransferase.

Science advances·2025

Same author

Impact of geothermal activity on the anomalous retreat of Changmolung Glacier in the Karakoram.

The Science of the total environment·2025

Same author

Humidity Assisted Selective Ion Exclusion in 0D Mixed Halide Perovskite Related Structures for Multimode Anti-counterfeiting.

Small (Weinheim an der Bergstrasse, Germany)·2025

Same author

Rim learning framework based on TS-GAN: A new paradigm of automated glaucoma screening from fundus images.

Computers in biology and medicine·2025

Same author

Development of multifunctional fluorescence-emitting potential theranostic agents for Alzheimer's disease.

Talanta·2025

Search research articles

Related Experiment Video

Updated: Jun 12, 2026

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

Published on: January 18, 2020

When does global attention help: a unified empirical study on atomistic graph learning.

Arindam Chowdhury¹, Massimiliano Lupo Pasini²

¹Computational Sciences and Engineering Division (CSED), Oak Ridge National Laboratory, 5700, 1 Bethel Valley Rd, Oak Ridge, TN, 37830, USA.

Journal of Cheminformatics

|May 1, 2026

Summary

Graph neural networks (GNNs) offer atomistic insights, but global attention benefits remain unclear. Our framework shows encoder-augmented MPNNs are robust, while fused models excel for long-range interactions.

Keywords:

Equivariant graph neural networks Graph transformer Long-range interactions Message passing neural network Topological and chemical encoders

More Related Videos

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Published on: June 13, 2025

Related Experiment Videos

Last Updated: Jun 12, 2026

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

A Methodology for Capturing Joint Visual Attention Using Mobile Eye-Trackers

Published on: January 18, 2020

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Author Spotlight: Addressing Technical and Subjective Challenges in Measuring Classroom Attention

Published on: December 15, 2023

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Evidence-based Knowledge Synthesis and Hypothesis Validation: Navigating Biomedical Knowledge Bases via Explainable AI and Agentic Systems

Published on: June 13, 2025

Area of Science:

Computational chemistry and materials science.
Machine learning and artificial intelligence.
Atomistic simulations and modeling.

Background:

Graph neural networks (GNNs) are crucial for atomistic simulations, but architectural complexity, especially global attention mechanisms, requires systematic evaluation.
Current GNN implementations often lack consistency, making it difficult to ascertain the true benefits of global attention over traditional message passing neural networks (MPNNs).

Purpose of the Study:

To introduce a unified, reproducible benchmarking framework (HydraGNN) for evaluating GNN architectures in atomistic learning.
To systematically isolate and quantify the contributions of message passing, global attention, and feature augmentation in GNNs.
To provide clear guidelines on when global attention mechanisms offer tangible advantages over MPNNs.

Main Methods:

Development of HydraGNN, a flexible framework enabling seamless switching between four GNN model classes: MPNN, MPNN with encoders, GPS-style hybrids, and fused local-global models.
Benchmarking across seven diverse datasets for regression and classification tasks, focusing on atomistic properties.
Systematic isolation of contributions from message passing, global attention, and encoder-based feature augmentation.

Main Results:

Encoder-augmented MPNNs establish a strong, reliable baseline performance.
Fused local-global GNN models demonstrate significant advantages for properties dominated by long-range interactions.
Quantification of the accuracy-compute trade-offs associated with global attention, including memory overhead.

Conclusions:

This study provides the first controlled evaluation of global attention in atomistic graph learning.
Fused local-global models are recommended for properties influenced by long-range interactions.
The HydraGNN framework serves as a reproducible testbed for future GNN model development in atomistic simulations.