Towards a Generalizable AI for Materials Discovery: Validation through Immersion Coolant Screening

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Hyunseung, Jeong, Dae-Woong, Park, Changyoung, Lee, Won-Ji, Lee, Ha-Eun, Lee, Ji-Hye, Hormazabal, Rodrigo, Ko, Sung Moon, Lee, Sumin, Yim, Soorin, Lee, Chanhui, Han, Sehui, Cha, Sang-Ho, Lim, Woohyung
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911242338172928
author Kim, Hyunseung
Jeong, Dae-Woong
Park, Changyoung
Lee, Won-Ji
Lee, Ha-Eun
Lee, Ji-Hye
Hormazabal, Rodrigo
Ko, Sung Moon
Lee, Sumin
Yim, Soorin
Lee, Chanhui
Han, Sehui
Cha, Sang-Ho
Lim, Woohyung
author_facet Kim, Hyunseung
Jeong, Dae-Woong
Park, Changyoung
Lee, Won-Ji
Lee, Ha-Eun
Lee, Ji-Hye
Hormazabal, Rodrigo
Ko, Sung Moon
Lee, Sumin
Yim, Soorin
Lee, Chanhui
Han, Sehui
Cha, Sang-Ho
Lim, Woohyung
contents Artificial intelligence (AI) has emerged as a powerful accelerator of materials discovery, yet most existing models remain problem-specific, requiring additional data collection and retraining for each new property. Here we introduce and validate GATE (Geometrically Aligned Transfer Encoder) -- a generalizable AI framework that jointly learns 34 physicochemical properties spanning thermal, electrical, mechanical, and optical domains. By aligning these properties within a shared geometric space, GATE captures cross-property correlations that reduce disjoint-property bias -- a key factor causing false positives in multi-criteria screening. To demonstrate its generalizable utility, GATE -- without any problem-specific model reconfiguration -- applied to the discovery of immersion cooling fluids for data centers, a stringent real-world challenge defined by the Open Compute Project (OCP). Screening billions of candidates, GATE identified 92,861 molecules as promising for practical deployment. Four were experimentally or literarily validated, showing strong agreement with wet-lab measurements and performance comparable to or exceeding a commercial coolant. These results establish GATE as a generalizable AI platform readily applicable across diverse materials discovery tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2510_23371
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards a Generalizable AI for Materials Discovery: Validation through Immersion Coolant Screening
Kim, Hyunseung
Jeong, Dae-Woong
Park, Changyoung
Lee, Won-Ji
Lee, Ha-Eun
Lee, Ji-Hye
Hormazabal, Rodrigo
Ko, Sung Moon
Lee, Sumin
Yim, Soorin
Lee, Chanhui
Han, Sehui
Cha, Sang-Ho
Lim, Woohyung
Machine Learning
Computational Engineering, Finance, and Science
Artificial intelligence (AI) has emerged as a powerful accelerator of materials discovery, yet most existing models remain problem-specific, requiring additional data collection and retraining for each new property. Here we introduce and validate GATE (Geometrically Aligned Transfer Encoder) -- a generalizable AI framework that jointly learns 34 physicochemical properties spanning thermal, electrical, mechanical, and optical domains. By aligning these properties within a shared geometric space, GATE captures cross-property correlations that reduce disjoint-property bias -- a key factor causing false positives in multi-criteria screening. To demonstrate its generalizable utility, GATE -- without any problem-specific model reconfiguration -- applied to the discovery of immersion cooling fluids for data centers, a stringent real-world challenge defined by the Open Compute Project (OCP). Screening billions of candidates, GATE identified 92,861 molecules as promising for practical deployment. Four were experimentally or literarily validated, showing strong agreement with wet-lab measurements and performance comparable to or exceeding a commercial coolant. These results establish GATE as a generalizable AI platform readily applicable across diverse materials discovery tasks.
title Towards a Generalizable AI for Materials Discovery: Validation through Immersion Coolant Screening
topic Machine Learning
Computational Engineering, Finance, and Science
url https://arxiv.org/abs/2510.23371