BEGIN:VCALENDAR VERSION:2.0 PRODID:-//chikkutakku.com//RDFCal 1.0//EN X-WR-CALDESC:GoogleカレンダーやiCalendar形式情報を共有シェ アしましょう。近所のイベントから全国のイベントま で今日のイベント検索やスケジュールを決めるならち っくたっく X-WR-CALNAME:ちっくたっく X-WR-TIMEZONE:UTC BEGIN:VEVENT SUMMARY:ChE Seminar: John Kitchin\, Carnegie Mellon University\, "Large ma chine-learned potentials for materials design with catalysis and energy ap plications" DTSTART;VALUE=DATE-TIME:20260407T153000Z DTEND;VALUE=DATE-TIME:20260407T163000Z UID:148344059178 DESCRIPTION: Sponsored by the Department of Chemical and Biomolecular Eng ineeringHost: AlexandraZagalskayaazagalskaya@umass.edu 413-545-7114John Ki tchin Carnegie Mellon University “Large machine-learnedpotentials for m aterials design with catalysis and energy applications” Tuesday\, April 7\, 2026\, 11:30 a.m.201 LGRT\, UMass Amherst(Refreshments at 11:15 a.m.)  Abstract:Designing multicomponent materials ischallenging with density f unctional theory (DFT) because it computationallyexpensive to evaluate all the ways that elements may combine and react. It isalso difficult to esti mate free energy contributions to reactions and to locatereaction barriers with DFT. The Open Catalyst Project is developing machinelearned potentia ls (MLP) to mitigate these challenges. These MLPs are trainedon 100M+ DFT calculations spanning 55 different elements and 80+ adsorbatesthat are rel evant in catalysis and energy applications. Nominally these modelswere tra ined to predict energy and forces\, and from these one can derivereaction energies. We will show in this talk\, however\, that these models alsoshow great utility in computing reaction barriers\, and in estimating freeener gy contributions to reactions. This opens the door to a post-scaling era o fcomputational catalysis where reaction barriers can be computed in comple xreaction networks with near DFT accuracy rather than relying on less accu ratelinear scaling relations. We will show some case studies of results an d discussfuture research directions in this area. Finally\, I will discuss the growingrole of generative AI in scientific research with examples fro m our most recentwork. Bio: John Kitchin completed his B.S. in Chemistry at North Carolina State University. He completed a M.S. in Materials Scien ceand a PhD in Chemical Engineering at the University of Delaware in 2004 underthe advisement of Dr. Jingguang Chen and Dr. Mark Barteau. He receive d anAlexander von Humboldt postdoctoral fellowship and lived in Berlin\, G ermany for1 ½ years studying alloy segregation with Karsten Reuter and Ma tthias Schefflerin the Theory Department at the Fritz Haber Institut. Prof essor Kitchin began atenure-track faculty position in the Chemical Enginee ring Department atCarnegie Mellon University in January of 2006. He is cur rently the John E.Swearingen Professor. At CMU\, Professor Kitchin works i n the areas of alloycatalysis and molecular simulation. He was awarded a D OE Early Career award in2010 to investigate multifunctional oxide electroc atalysts for the oxygenevolution reaction in water splitting using experim ental and computationalmethods. He received a Presidential Early Career Aw ard for Scientists andEngineers in 2011. He completed a sabbatical in the Accelerated Science groupat Google learning to apply machine learning to s cientific and engineeringproblems in 2018. LOCATION:LGRT 201 END:VEVENT END:VCALENDAR