{"657009":{"#nid":"657009","#data":{"type":"event","title":"Physical Chemistry Seminar:  Prof.  Jason Goodpaster","body":[{"value":"\u003Cp\u003EAbstract: Large, periodic, and condensed phase systems impose a challenge for theoretical studies due to the compromise between accuracy and computational cost in their\u0026nbsp;calculations. \u0026nbsp;We present two methods that show exciting promise for treating this compromise: machine learning and quantum embedding. We exploit machine learning\u0026nbsp;methods to solve this accuracy and computational cost trade-off by leveraging large data sets to train on highly accurate calculations using small molecules and then apply\u0026nbsp;them to larger systems. We are developing a method to train a neural network potential with high-level wavefunction theory on targeted systems of interest that can describe\u0026nbsp;bond breaking. We combine density functional theory calculations and higher-level ab initio wavefunction calculations, such as CASPT2, to train our neural network\u0026nbsp;potentials. We first train our neural network at the DFT level of theory. Using an adaptive active learning training scheme, we retrained the neural network potential to a\u0026nbsp;CASPT2 level of accuracy. Quantum embedding methodology exploits the locality of chemical interactions to allow for accurate yet computationally efficient calculations to be\u0026nbsp;performed on complex systems. Quantum embedding allows for the partitioning of the system into two regions. \u0026nbsp;One is treated at a highly accurate level of theory using wave\u0026nbsp;function theory methods, and the other is treated at the more computationally efficient level of DFT. \u0026nbsp;We discuss our recent advancements for quantum embedding,\u0026nbsp;specifically for systems with complicated electronic structure such as metal organic frameworks. \u0026nbsp;Together, we believe both methodologies can allow for complex systems to\u0026nbsp;be studied at a significantly reduced computational cost.\u003C\/p\u003E\r\n","summary":null,"format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003ELarge, periodic, and condensed phase systems impose a challenge for theoretical studies due to the compromise between accuracy and computational cost in their\u0026nbsp;calculations\u003C\/p\u003E\r\n","format":"limited_html"}],"field_summary_sentence":[{"value":"Advancements in Machine Learning and Quantum Embedding for Large Scale Simulations"}],"uid":"34592","created_gmt":"2022-04-05 19:42:26","changed_gmt":"2022-04-05 19:42:26","author":"wh105","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2022-04-12T17:00:00-04:00","event_time_end":"2022-04-12T18:00:00-04:00","event_time_end_last":"2022-04-12T18:00:00-04:00","gmt_time_start":"2022-04-12 21:00:00","gmt_time_end":"2022-04-12 22:00:00","gmt_time_end_last":"2022-04-12 22:00:00","rrule":null,"timezone":"America\/New_York"},"extras":[],"related_links":[{"url":"https:\/\/bluejeans.com\/764447847","title":""}],"groups":[{"id":"85951","name":"School of Chemistry and Biochemistry"}],"categories":[],"keywords":[{"id":"190316","name":"Machine Learning and Quantum Embedding"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1795","name":"Seminar\/Lecture\/Colloquium"}],"invited_audience":[{"id":"78761","name":"Faculty\/Staff"},{"id":"78771","name":"Public"},{"id":"78751","name":"Undergraduate students"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[{"value":"\u003Cp\u003EProf. Joshua Kretchmer, School of Chemistry and Biochemistry\u003C\/p\u003E\r\n","format":"limited_html"}],"email":[],"slides":[],"orientation":[],"userdata":""}}}