Loading…
Open Challenges in Developing Generalizable Large-Scale Machine-Learning Models for Catalyst Discovery
The development of machine-learned potentials for catalyst discovery has predominantly been focused on very specific chemistries and material compositions. While they are effective in interpolating between available materials, these approaches struggle to generalize across chemical space. The recent...
Saved in:
Published in: | ACS catalysis 2022-07, Vol.12 (14), p.8572-8581 |
---|---|
Main Authors: | , , , , , , , , |
Format: | Article |
Language: | English |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
cited_by | cdi_FETCH-LOGICAL-a280t-128357cc51bdcb145811cc97a2f3b72e1eb904f0214a5c59d85644dd4cda952f3 |
---|---|
cites | cdi_FETCH-LOGICAL-a280t-128357cc51bdcb145811cc97a2f3b72e1eb904f0214a5c59d85644dd4cda952f3 |
container_end_page | 8581 |
container_issue | 14 |
container_start_page | 8572 |
container_title | ACS catalysis |
container_volume | 12 |
creator | Kolluru, Adeesh Shuaibi, Muhammed Palizhati, Aini Shoghi, Nima Das, Abhishek Wood, Brandon Zitnick, C. Lawrence Kitchin, John R. Ulissi, Zachary W. |
description | The development of machine-learned potentials for catalyst discovery has predominantly been focused on very specific chemistries and material compositions. While they are effective in interpolating between available materials, these approaches struggle to generalize across chemical space. The recent curation of large-scale catalyst data sets has offered the opportunity to build a universal machine-learning potential, spanning chemical and composition space. If accomplished, said potential could accelerate the catalyst discovery process across a variety of applications (CO2 reduction, NH3 production, etc.) without the additional specialized training efforts that are currently required. The release of the Open Catalyst 2020 Data set (OC20) has begun just that, pushing the heterogeneous catalysis and machine-learning communities toward building more accurate and robust models. In this Perspective, we discuss some of the challenges and findings of recent developments on OC20. We examine the performance of current models across different materials and adsorbates to identify notably underperforming subsets. We then discuss some of the modeling efforts surrounding energy conservation, approaches to finding and evaluating the local minima, and augmentation of off-equilibrium data. To complement the community’s ongoing developments, we end with an outlook to some of the important challenges that have yet to be thoroughly explored for large-scale catalyst discovery. |
doi_str_mv | 10.1021/acscatal.2c02291 |
format | article |
fullrecord | <record><control><sourceid>acs_cross</sourceid><recordid>TN_cdi_crossref_primary_10_1021_acscatal_2c02291</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>a98709052</sourcerecordid><originalsourceid>FETCH-LOGICAL-a280t-128357cc51bdcb145811cc97a2f3b72e1eb904f0214a5c59d85644dd4cda952f3</originalsourceid><addsrcrecordid>eNp1UMFOAjEQbYwmEuTusR_gYltadvdoFgWTJRzU82Z2dhaW1C5pkQS_3hIw8eJc5iXz3uS9x9i9FGMplHwEDAh7sGOFQqlcXrGBksYkRk_M9R98y0YhbEUcbaZZKgasXe3I8WID1pJbU-Cd4zM6kO13nVvzOTnyYLtvqC3xEvyakjeEiJeAm85RUhJ4d6Iu-4Zs4G3veXHycgx7PusC9gfyxzt204INNLrsIft4eX4vFkm5mr8WT2UCKhP7RKpsYlJEI-sGa6lNJiVinoJqJ3WqSFKdC93GyBoMmrzJzFTrptHYQG4iacjE-S_6PgRPbbXz3Sf4YyVFdaqq-q2qulQVJQ9nSbxU2_7Lu2jwf_oPBG5uBw</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Open Challenges in Developing Generalizable Large-Scale Machine-Learning Models for Catalyst Discovery</title><source>American Chemical Society:Jisc Collections:American Chemical Society Read & Publish Agreement 2022-2024 (Reading list)</source><creator>Kolluru, Adeesh ; Shuaibi, Muhammed ; Palizhati, Aini ; Shoghi, Nima ; Das, Abhishek ; Wood, Brandon ; Zitnick, C. Lawrence ; Kitchin, John R. ; Ulissi, Zachary W.</creator><creatorcontrib>Kolluru, Adeesh ; Shuaibi, Muhammed ; Palizhati, Aini ; Shoghi, Nima ; Das, Abhishek ; Wood, Brandon ; Zitnick, C. Lawrence ; Kitchin, John R. ; Ulissi, Zachary W.</creatorcontrib><description>The development of machine-learned potentials for catalyst discovery has predominantly been focused on very specific chemistries and material compositions. While they are effective in interpolating between available materials, these approaches struggle to generalize across chemical space. The recent curation of large-scale catalyst data sets has offered the opportunity to build a universal machine-learning potential, spanning chemical and composition space. If accomplished, said potential could accelerate the catalyst discovery process across a variety of applications (CO2 reduction, NH3 production, etc.) without the additional specialized training efforts that are currently required. The release of the Open Catalyst 2020 Data set (OC20) has begun just that, pushing the heterogeneous catalysis and machine-learning communities toward building more accurate and robust models. In this Perspective, we discuss some of the challenges and findings of recent developments on OC20. We examine the performance of current models across different materials and adsorbates to identify notably underperforming subsets. We then discuss some of the modeling efforts surrounding energy conservation, approaches to finding and evaluating the local minima, and augmentation of off-equilibrium data. To complement the community’s ongoing developments, we end with an outlook to some of the important challenges that have yet to be thoroughly explored for large-scale catalyst discovery.</description><identifier>ISSN: 2155-5435</identifier><identifier>EISSN: 2155-5435</identifier><identifier>DOI: 10.1021/acscatal.2c02291</identifier><language>eng</language><publisher>American Chemical Society</publisher><ispartof>ACS catalysis, 2022-07, Vol.12 (14), p.8572-8581</ispartof><rights>2022 American Chemical Society</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-a280t-128357cc51bdcb145811cc97a2f3b72e1eb904f0214a5c59d85644dd4cda952f3</citedby><cites>FETCH-LOGICAL-a280t-128357cc51bdcb145811cc97a2f3b72e1eb904f0214a5c59d85644dd4cda952f3</cites><orcidid>0000-0002-9401-4918 ; 0000-0003-2625-9232</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,27924,27925</link.rule.ids></links><search><creatorcontrib>Kolluru, Adeesh</creatorcontrib><creatorcontrib>Shuaibi, Muhammed</creatorcontrib><creatorcontrib>Palizhati, Aini</creatorcontrib><creatorcontrib>Shoghi, Nima</creatorcontrib><creatorcontrib>Das, Abhishek</creatorcontrib><creatorcontrib>Wood, Brandon</creatorcontrib><creatorcontrib>Zitnick, C. Lawrence</creatorcontrib><creatorcontrib>Kitchin, John R.</creatorcontrib><creatorcontrib>Ulissi, Zachary W.</creatorcontrib><title>Open Challenges in Developing Generalizable Large-Scale Machine-Learning Models for Catalyst Discovery</title><title>ACS catalysis</title><addtitle>ACS Catal</addtitle><description>The development of machine-learned potentials for catalyst discovery has predominantly been focused on very specific chemistries and material compositions. While they are effective in interpolating between available materials, these approaches struggle to generalize across chemical space. The recent curation of large-scale catalyst data sets has offered the opportunity to build a universal machine-learning potential, spanning chemical and composition space. If accomplished, said potential could accelerate the catalyst discovery process across a variety of applications (CO2 reduction, NH3 production, etc.) without the additional specialized training efforts that are currently required. The release of the Open Catalyst 2020 Data set (OC20) has begun just that, pushing the heterogeneous catalysis and machine-learning communities toward building more accurate and robust models. In this Perspective, we discuss some of the challenges and findings of recent developments on OC20. We examine the performance of current models across different materials and adsorbates to identify notably underperforming subsets. We then discuss some of the modeling efforts surrounding energy conservation, approaches to finding and evaluating the local minima, and augmentation of off-equilibrium data. To complement the community’s ongoing developments, we end with an outlook to some of the important challenges that have yet to be thoroughly explored for large-scale catalyst discovery.</description><issn>2155-5435</issn><issn>2155-5435</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><recordid>eNp1UMFOAjEQbYwmEuTusR_gYltadvdoFgWTJRzU82Z2dhaW1C5pkQS_3hIw8eJc5iXz3uS9x9i9FGMplHwEDAh7sGOFQqlcXrGBksYkRk_M9R98y0YhbEUcbaZZKgasXe3I8WID1pJbU-Cd4zM6kO13nVvzOTnyYLtvqC3xEvyakjeEiJeAm85RUhJ4d6Iu-4Zs4G3veXHycgx7PusC9gfyxzt204INNLrsIft4eX4vFkm5mr8WT2UCKhP7RKpsYlJEI-sGa6lNJiVinoJqJ3WqSFKdC93GyBoMmrzJzFTrptHYQG4iacjE-S_6PgRPbbXz3Sf4YyVFdaqq-q2qulQVJQ9nSbxU2_7Lu2jwf_oPBG5uBw</recordid><startdate>20220715</startdate><enddate>20220715</enddate><creator>Kolluru, Adeesh</creator><creator>Shuaibi, Muhammed</creator><creator>Palizhati, Aini</creator><creator>Shoghi, Nima</creator><creator>Das, Abhishek</creator><creator>Wood, Brandon</creator><creator>Zitnick, C. Lawrence</creator><creator>Kitchin, John R.</creator><creator>Ulissi, Zachary W.</creator><general>American Chemical Society</general><scope>AAYXX</scope><scope>CITATION</scope><orcidid>https://orcid.org/0000-0002-9401-4918</orcidid><orcidid>https://orcid.org/0000-0003-2625-9232</orcidid></search><sort><creationdate>20220715</creationdate><title>Open Challenges in Developing Generalizable Large-Scale Machine-Learning Models for Catalyst Discovery</title><author>Kolluru, Adeesh ; Shuaibi, Muhammed ; Palizhati, Aini ; Shoghi, Nima ; Das, Abhishek ; Wood, Brandon ; Zitnick, C. Lawrence ; Kitchin, John R. ; Ulissi, Zachary W.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a280t-128357cc51bdcb145811cc97a2f3b72e1eb904f0214a5c59d85644dd4cda952f3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Kolluru, Adeesh</creatorcontrib><creatorcontrib>Shuaibi, Muhammed</creatorcontrib><creatorcontrib>Palizhati, Aini</creatorcontrib><creatorcontrib>Shoghi, Nima</creatorcontrib><creatorcontrib>Das, Abhishek</creatorcontrib><creatorcontrib>Wood, Brandon</creatorcontrib><creatorcontrib>Zitnick, C. Lawrence</creatorcontrib><creatorcontrib>Kitchin, John R.</creatorcontrib><creatorcontrib>Ulissi, Zachary W.</creatorcontrib><collection>CrossRef</collection><jtitle>ACS catalysis</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Kolluru, Adeesh</au><au>Shuaibi, Muhammed</au><au>Palizhati, Aini</au><au>Shoghi, Nima</au><au>Das, Abhishek</au><au>Wood, Brandon</au><au>Zitnick, C. Lawrence</au><au>Kitchin, John R.</au><au>Ulissi, Zachary W.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Open Challenges in Developing Generalizable Large-Scale Machine-Learning Models for Catalyst Discovery</atitle><jtitle>ACS catalysis</jtitle><addtitle>ACS Catal</addtitle><date>2022-07-15</date><risdate>2022</risdate><volume>12</volume><issue>14</issue><spage>8572</spage><epage>8581</epage><pages>8572-8581</pages><issn>2155-5435</issn><eissn>2155-5435</eissn><abstract>The development of machine-learned potentials for catalyst discovery has predominantly been focused on very specific chemistries and material compositions. While they are effective in interpolating between available materials, these approaches struggle to generalize across chemical space. The recent curation of large-scale catalyst data sets has offered the opportunity to build a universal machine-learning potential, spanning chemical and composition space. If accomplished, said potential could accelerate the catalyst discovery process across a variety of applications (CO2 reduction, NH3 production, etc.) without the additional specialized training efforts that are currently required. The release of the Open Catalyst 2020 Data set (OC20) has begun just that, pushing the heterogeneous catalysis and machine-learning communities toward building more accurate and robust models. In this Perspective, we discuss some of the challenges and findings of recent developments on OC20. We examine the performance of current models across different materials and adsorbates to identify notably underperforming subsets. We then discuss some of the modeling efforts surrounding energy conservation, approaches to finding and evaluating the local minima, and augmentation of off-equilibrium data. To complement the community’s ongoing developments, we end with an outlook to some of the important challenges that have yet to be thoroughly explored for large-scale catalyst discovery.</abstract><pub>American Chemical Society</pub><doi>10.1021/acscatal.2c02291</doi><tpages>10</tpages><orcidid>https://orcid.org/0000-0002-9401-4918</orcidid><orcidid>https://orcid.org/0000-0003-2625-9232</orcidid></addata></record> |
fulltext | fulltext |
identifier | ISSN: 2155-5435 |
ispartof | ACS catalysis, 2022-07, Vol.12 (14), p.8572-8581 |
issn | 2155-5435 2155-5435 |
language | eng |
recordid | cdi_crossref_primary_10_1021_acscatal_2c02291 |
source | American Chemical Society:Jisc Collections:American Chemical Society Read & Publish Agreement 2022-2024 (Reading list) |
title | Open Challenges in Developing Generalizable Large-Scale Machine-Learning Models for Catalyst Discovery |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T14%3A24%3A11IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-acs_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Open%20Challenges%20in%20Developing%20Generalizable%20Large-Scale%20Machine-Learning%20Models%20for%20Catalyst%20Discovery&rft.jtitle=ACS%20catalysis&rft.au=Kolluru,%20Adeesh&rft.date=2022-07-15&rft.volume=12&rft.issue=14&rft.spage=8572&rft.epage=8581&rft.pages=8572-8581&rft.issn=2155-5435&rft.eissn=2155-5435&rft_id=info:doi/10.1021/acscatal.2c02291&rft_dat=%3Cacs_cross%3Ea98709052%3C/acs_cross%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-a280t-128357cc51bdcb145811cc97a2f3b72e1eb904f0214a5c59d85644dd4cda952f3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true |