{"690021":{"#nid":"690021","#data":{"type":"event","title":"PhD Defense by Zilong Wang","body":[{"value":"\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003ETitle: Applications of Optimization and Causal Inference\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EDate: Wednesday, May 6, 2026\u003C\/p\u003E\u003Cp\u003ETime: 3:00 pm \u2013 5:00 pm ET\u003C\/p\u003E\u003Cp\u003ELocation: Remote on Zoom \u2014 \u003Ca href=\u0022https:\/\/nam12.safelinks.protection.outlook.com\/?url=https%3A%2F%2Fgatech.zoom.us%2Fj%2F93388746171%3Fpwd%3DIknDRQHq8fIL8QJi7YQTV9kveeYnxp.1\u0026amp;data=05%7C02%7Ctm186%40gtvault.onmicrosoft.com%7Cddc37aace4634964f88c08dea44dab71%7C482198bbae7b4b258b7a6d7f32faa083%7C1%7C0%7C639128850683090519%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C\u0026amp;sdata=gTtMhNY6zp0rgcPU2wxG1ZoFjhfiVj6ZL5%2FQb240mOM%3D\u0026amp;reserved=0\u0022\u003Ehttps:\/\/gatech.zoom.us\/j\/93388746171?pwd=IknDRQHq8fIL8QJi7YQTV9kveeYnxp.1\u003C\/a\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EMeeting ID: 933 8874 6171\u003C\/p\u003E\u003Cp\u003EPasscode: 734561\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003ECommittee:\u003C\/p\u003E\u003Cp\u003EDr. Turgay Ayer (Advisor), Virginia C. and Joseph C. Mello Endowed Chair, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology\u003C\/p\u003E\u003Cp\u003EDr. Shihao Yang, Harold E. Smalley Early Career Professor, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology\u003C\/p\u003E\u003Cp\u003EDr. Santanu Dey, Anderson-Interface Chair and Professor, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology\u003C\/p\u003E\u003Cp\u003EDr. David Goldsman, Coca-Cola Foundation Professor, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology\u003C\/p\u003E\u003Cp\u003EDr. Jagpreet Chhatwal, Director, Center for Health Technology Assessment, Mass General Brigham; Associate Professor, Harvard Medical School\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EAbstract:\u003C\/p\u003E\u003Cp\u003EThis thesis develops methods for causal learning under\u003C\/p\u003E\u003Cp\u003Edata scarcity and distributional shift, combining transfer learning,\u003C\/p\u003E\u003Cp\u003Ekernel learning, and doubly robust semi-parametric estimation in public health\u0026nbsp;\u003C\/p\u003E\u003Cp\u003Eand personalized medicine.\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe first part develops Transfer Learning Random Forests (TLRF) for\u003C\/p\u003E\u003Cp\u003Ecounty-level epidemic growth-rate estimation and outbreak prioritization\u003C\/p\u003E\u003Cp\u003Eunder fixed public-health capacity. Its core technical step is a\u003C\/p\u003E\u003Cp\u003Eproxy-identification argument that recasts latent\u003C\/p\u003E\u003Cp\u003Egrowth-rate estimation as conditional mean learning on observable\u003C\/p\u003E\u003Cp\u003Elog-difference outcomes, followed by adaptive spatiotemporal transfer\u003C\/p\u003E\u003Cp\u003Evia random-forest weighting. In a retrospective comparison against\u003C\/p\u003E\u003Cp\u003Ehistorical decisions by the Colorado Department of Public Health and\u003C\/p\u003E\u003Cp\u003EEnvironment, TLRF yields substantially higher investigation positive\u003C\/p\u003E\u003Cp\u003Epredictive value (including a 224% relative improvement as a point\u003C\/p\u003E\u003Cp\u003Eestimate under the same decision budget).\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe second part develops a causal-kernel subgroup discovery framework:\u003C\/p\u003E\u003Cp\u003ENeyman-orthogonalized CATE estimation and kernel learning under the\u003C\/p\u003E\u003Cp\u003ERobinson decomposition are followed by kernelized convex clustering to\u003C\/p\u003E\u003Cp\u003Eidentify subpopulations defined by treatment responsiveness rather than\u003C\/p\u003E\u003Cp\u003Ecovariate similarity alone. This component turns treatment-effect\u003C\/p\u003E\u003Cp\u003Eheterogeneity into decision-ready subgroup structure to support\u003C\/p\u003E\u003Cp\u003Ehypothesis generation, follow-up trial enrichment, and targeted\u003C\/p\u003E\u003Cp\u003Eintervention design.\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EThe third part develops a placebo-anchored transfer framework for\u003C\/p\u003E\u003Cp\u003Etransporting individualized treatment effects across heterogeneous\u003C\/p\u003E\u003Cp\u003Etrials under covariate shift, including settings where the target\u003C\/p\u003E\u003Cp\u003Epopulation has limited or no treated outcomes. A sparsity-driven,\u003C\/p\u003E\u003Cp\u003ELASSO-based proxy-gold transfer correction is integrated with\u003C\/p\u003E\u003Cp\u003Ecross-fitted doubly robust estimation to support target-population\u003C\/p\u003E\u003Cp\u003Eeffect learning under explicit transport assumptions. This provides a\u003C\/p\u003E\u003Cp\u003Epractical evidence-synthesis pathway for target-population decision\u003C\/p\u003E\u003Cp\u003Esupport when standard IPD meta-analysis assumptions are not fully met.\u003C\/p\u003E\u003Cp\u003E\u0026nbsp;\u003C\/p\u003E\u003Cp\u003EAcross these three settings, the unifying theme is principled\u003C\/p\u003E\u003Cp\u003Einformation borrowing: semiparametric identification, orthogonalized\u003C\/p\u003E\u003Cp\u003Ecausal kernel learning, and transfer-based doubly robust inference for\u003C\/p\u003E\u003Cp\u003Edecision-making when target data alone are insufficient.\u003C\/p\u003E","summary":"","format":"limited_html"}],"field_subtitle":"","field_summary":[{"value":"\u003Cp\u003EApplications of Optimization and Causal Inference\u003C\/p\u003E","format":"limited_html"}],"field_summary_sentence":[{"value":"Applications of Optimization and Causal Inference"}],"uid":"27707","created_gmt":"2026-04-27 12:12:17","changed_gmt":"2026-04-27 12:12:47","author":"Tatianna Richardson","boilerplate_text":"","field_publication":"","field_article_url":"","field_event_time":{"event_time_start":"2026-05-06T15:00:00-04:00","event_time_end":"2026-05-06T17:00:00-04:00","event_time_end_last":"2026-05-06T17:00:00-04:00","gmt_time_start":"2026-05-06 19:00:00","gmt_time_end":"2026-05-06 21:00:00","gmt_time_end_last":"2026-05-06 21:00:00","rrule":null,"timezone":"America\/New_York"},"location":"ZOOM","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":""}}}