Related Experiment Video
Updated: Jul 26, 2025

Author Spotlight: Streamlining Visual Dynamics to Simplify Molecular Dynamics Simulations Using Gromacs
Published on: August 9, 2024
Updates to the DScribe library: New descriptors and derivatives
Jarno Laakso1, Lauri Himanen1, Henrietta Homm1
1Department of Applied Physics, Aalto University, P.O. Box 11100, 00076 Aalto, Finland.
The DScribe Python package now offers new materials fingerprints and descriptor derivatives for advanced machine learning, enhancing atomistic simulations and materials discovery.
Area of Science:
- Computational materials science
- Machine learning for chemistry
- Python software development
Background:
- Atomistic descriptors are crucial for machine learning models in materials science.
- Existing descriptor packages may lack advanced features for complex tasks.
- The DScribe package facilitates the use of atomistic descriptors in Python.
Purpose of the Study:
- To update the DScribe Python package with new functionalities.
- To enable advanced machine learning tasks like force prediction and structure optimization.
- To improve the accuracy and applicability of atomistic descriptors.
Main Methods:
- Integration of the Valle-Oganov materials fingerprint.
- Implementation of numeric and analytic descriptor derivatives.
- Development of derivatives for Many-Body Tensor Representation (MBTR) and Smooth Overlap of Atomic Positions (SOAP).
Main Results:
- DScribe now includes the Valle-Oganov fingerprint.
- Numeric derivatives are available for all descriptors.
- Analytic derivatives are implemented for MBTR and SOAP descriptors.
- Demonstrated effectiveness of derivatives on Cu clusters and perovskite alloys.
Conclusions:
- The updated DScribe package enhances machine learning capabilities in materials science.
- Descriptor derivatives facilitate advanced predictive modeling and optimization.
- The new features support more sophisticated atomistic simulations and materials design.
Related Concept Videos
Doppler Effect - II
Diels–Alder Reaction: Characteristics of Dienes
Characteristics of the diene
Conformation
The simplest example of a diene is 1,3-butadiene, an acyclic conjugated π system. At room temperature, the molecule exists as a mixture of s-cis and s-trans conformers by virtue of rotation around the carbon–carbon single bond. Although the s-trans isomer is...
Stability of Conjugated Dienes
A comparison of the enthalpies of hydrogenation of dienes reveals that conjugated dienes release less heat on hydrogenation, rendering them more stable than their nonconjugated analogs.
Doppler Effect - I
Structure of Conjugated Dienes
Conjugated dienes are compounds characterized by the presence of alternating double and single bonds. In a conjugated system like 1,3-butadiene, the unhybridized 2p orbital on each carbon overlaps continuously, allowing the π electrons to be delocalized across the entire molecule. In contrast, this type of overlap does not occur in cumulated and isolated dienes, such as 2,3-pentadiene and 1,4-pentadiene, respectively. Instead, the π electrons remain localized between the double...
¹³C NMR: Distortionless Enhancement by Polarization Transfer (DEPT)

