Pith. sign in

REVIEW

Generative deep learning for the inverse design of materials

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2409.19124 v1 pith:B3I27C7F submitted 2024-09-27 cond-mat.mtrl-sci physics.comp-ph

Generative deep learning for the inverse design of materials

classification cond-mat.mtrl-sci physics.comp-ph
keywords learninggenerativematerialsdeeppropertycrystaldesigninverse
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
Share X Bluesky LinkedIn Reddit HN
read the original abstract

In addition to the forward inference of materials properties using machine learning, generative deep learning techniques applied on materials science allow the inverse design of materials, i.e., assessing the composition-processing-(micro-)structure-property relationships in a reversed way. In this review, we focus on the (micro-)structure-property mapping, i.e., crystal structure-intrinsic property and microstructure-extrinsic property, and summarize comprehensively how generative deep learning can be performed. Three key elements, i.e., the construction of latent spaces for both the crystal structures and microstructures, generative learning approaches, and property constraints, are discussed in detail. A perspective is given outlining the challenges of the existing methods in terms of computational resource consumption, data compatibility, and yield of generation.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.