{"680349":{"#nid":"680349","#data":{"type":"event","title":"PhD Defense by Matt Repasky","body":[{"value":"\u003Cp\u003E\u003Cstrong\u003ETitle:\u003C\/strong\u003E\u0026nbsp;Deep Learning for High-Dimensional Decision Making and Uncertainty Quantification\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EDate:\u0026nbsp;\u003C\/strong\u003EFebruary 24, 2025\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ETime:\u003C\/strong\u003E\u0026nbsp;1:00 PM - 2:30 PM ET\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ELocation:\u003C\/strong\u003E\u0026nbsp;Groseclose 402\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EZoom:\u003C\/strong\u003E\u0026nbsp;\u003Ca href=\u0022https:\/\/gatech.zoom.us\/j\/93830976845\u0022\u003Ehttps:\/\/gatech.zoom.us\/j\/93830976845\u003C\/a\u003E\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EMatthew Repasky\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003EMachine Learning PhD Student\u003C\/p\u003E\u003Cp\u003EH. Milton Stewart School of Industrial and Systems Engineering\u003C\/p\u003E\u003Cp\u003EGeorgia Institute of Technology\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003ECommittee\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003E1. Dr. Yao Xie (Advisor)\u003C\/p\u003E\u003Cp\u003E2. Dr. He Wang\u003C\/p\u003E\u003Cp\u003E3. Dr. Kamran Paynabar\u003C\/p\u003E\u003Cp\u003E4. Dr. Xiuyuan Cheng\u003C\/p\u003E\u003Cp\u003E5. Dr. Erwan Mazarico\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u003Cstrong\u003EAbstract\u003C\/strong\u003E\u003C\/p\u003E\u003Cp\u003EDeep learning facilitates solving problems of increasingly high dimension and complexity. This dissertation focuses on applying deep neural networks for tasks in decision making and uncertainty quantification. Single-step and sequential decision making is determined based on the network output, with specific applications to patrol and dispatch reinforcement learning and model critique. Neural model critics enable localization of disparity in model probability distributions, quantifying their goodness-of-fit. Similarly, robust samplers from probability models can aid in the characterization of prediction uncertainty. A method for solving Bayesian inverse problems using pre-trained generative models is defined, which effectively quantifies uncertainty in the posterior. Similarly, generative models are trained to learn the distribution of outputs for a lunar topography super-resolution task. The contributions outlined in this dissertation highlight the promise of neural networks for solving previously intractable problems while also providing uncertainty quantification.\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EDeep Learning for High-Dimensional Decision Making and Uncertainty Quantification\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Deep Learning for High-Dimensional Decision Making and Uncertainty Quantification"}],"uid":"27707","created_gmt":"2025-02-10 17:29:55","changed_gmt":"2025-02-10 17:31:29","author":"Tatianna Richardson","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2025-02-24T13:00:00-05:00","event_time_end":"2025-02-24T14:30:00-05:00","event_time_end_last":"2025-02-24T14:30:00-05:00","gmt_time_start":"2025-02-24 18:00:00","gmt_time_end":"2025-02-24 19:30:00","gmt_time_end_last":"2025-02-24 19:30:00","rrule":null,"timezone":"America\/New_York"},"location":"Groseclose 402","extras":[],"groups":[{"id":"221981","name":"Graduate Studies"}],"categories":[],"keywords":[{"id":"100811","name":"Phd Defense"}],"core_research_areas":[],"news_room_topics":[],"event_categories":[{"id":"1788","name":"Other\/Miscellaneous"}],"invited_audience":[{"id":"78771","name":"Public"}],"affiliations":[],"classification":[],"areas_of_expertise":[],"news_and_recent_appearances":[],"phone":[],"contact":[],"email":[],"slides":[],"orientation":[],"userdata":""}}}