BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Department of Statistics - ECPv6.15.20//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-ORIGINAL-URL:https://statistics.sciences.ncsu.edu
X-WR-CALDESC:Events for Department of Statistics
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:America/New_York
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20240310T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20241103T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20250309T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20251102T060000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
TZNAME:EDT
DTSTART:20260308T070000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
TZNAME:EST
DTSTART:20261101T060000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20251003T110000
DTEND;TZID=America/New_York:20251003T120000
DTSTAMP:20260504T094058
CREATED:20250922T182300Z
LAST-MODIFIED:20250922T182705Z
UID:28629-1759489200-1759492800@statistics.sciences.ncsu.edu
SUMMARY:Statistics Seminar
DESCRIPTION:Location: 2203 SAS Hall\, NC State Main Campus \nSpeaker: Elynn Chen \nAssistant Professor of Technology\, Operations and Statistics (TOPS) at NYU Stern School of Business \nTitle: Transfer Q-Learning: Stationary and Non-Stationary MDPs \nAbstract:\nIn dynamic decision-making scenarios across business\, healthcare\, and education\, leveraging data from diverse populations can significantly enhance reinforcement learning (RL) performance for specific target populations\, especially when target samples are limited. We develop comprehensive frameworks for transfer learning in RL\, addressing both stationary Markov decision processes (MDPs) with iterative Q-learning and non-stationary finite-horizon MDPs with backward inductive learning. \nFor stationary MDPs\, we propose an iterative Q-learning algorithm with knowledge transfer\, establishing theoretical justifications through faster convergence rates under similarity assumptions. For time-inhomogeneous finite-horizon MDPs\, we introduce two key innovations: (1) a novel “re-weighted targeting procedure” that enables vertical information-cascading along multiple temporal steps\, and (2) transfer deep Q-learning that leverages neural networks as function approximators. We demonstrate that while naive sample pooling strategies may succeed in regression settings\, they fail in MDPs\, necessitating our more sophisticated approach. We establish theoretical guarantees for both settings\, revealing the relationship between statistical performance and MDP task discrepancy. Our analysis illuminates how source and target sample sizes impact transfer effectiveness. The framework accommodates both transferable and non-transferable transition density ratios while assuming reward function transferability. Our analytical techniques have broader implications\, extending to supervised transfer learning with neural networks and domain shift scenarios. Empirical evidence from both synthetic and real datasets validates our theoretical results\, demonstrating significant improvements over single-task learning rates and highlighting the practical value of strategically constructed transferable RL samples in both stationary and non-stationary contexts.
URL:https://statistics.sciences.ncsu.edu/event/statistics-seminar-37/
LOCATION:NC
CATEGORIES:College of Sciences Calendar,Department,Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20251010T110000
DTEND;TZID=America/New_York:20251010T120000
DTSTAMP:20260504T094058
CREATED:20250922T183145Z
LAST-MODIFIED:20250922T183145Z
UID:28632-1760094000-1760097600@statistics.sciences.ncsu.edu
SUMMARY:Statistics Seminar
DESCRIPTION:Location: 2203 SAS Hall\, NC State Main Campus \nSpeaker:  Josh Startmer\,  Founder of StatQuest \nTitle: \nStatQuest: Origins plus musings on the intersection of Statistics and Machine Learning. \nABSTRACT: Although closely related\, subtle but important differences separate machine learning practitioners from statisticians. In this talk\, we will use statistical linear models to highlight these differences. Then\, we will show how to overcome them by combining linear models with regularization\, a machine learning method. The end result gives us the best of both statistics and machine learning in the form of a model that allows us to investigate mechanisms while being amenable to big datasets.
URL:https://statistics.sciences.ncsu.edu/event/statistics-seminar-38/
LOCATION:NC
CATEGORIES:College of Sciences Calendar,Department,Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;VALUE=DATE:20251013
DTEND;VALUE=DATE:20251014
DTSTAMP:20260504T094058
CREATED:20250203T180349Z
LAST-MODIFIED:20250203T181954Z
UID:27774-1760313600-1760399999@statistics.sciences.ncsu.edu
SUMMARY:No Classes - Fall Break
DESCRIPTION:
URL:https://statistics.sciences.ncsu.edu/event/no-classes-fall-break-3/
LOCATION:NC
CATEGORIES:Department,Faculty,Graduate,Undergraduate,University
END:VEVENT
BEGIN:VEVENT
DTSTART;VALUE=DATE:20251013
DTEND;VALUE=DATE:20251014
DTSTAMP:20260504T094058
CREATED:20250203T180349Z
LAST-MODIFIED:20250203T181954Z
UID:27774-1760313600-1760399999@statistics.sciences.ncsu.edu
SUMMARY:No Classes - Fall Break
DESCRIPTION:
URL:https://statistics.sciences.ncsu.edu/event/no-classes-fall-break-3/
LOCATION:NC
CATEGORIES:Department,Faculty,Graduate,Undergraduate,University
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20251024T110000
DTEND;TZID=America/New_York:20251024T120000
DTSTAMP:20260504T094058
CREATED:20251013T174330Z
LAST-MODIFIED:20251013T174330Z
UID:28836-1761303600-1761307200@statistics.sciences.ncsu.edu
SUMMARY:Statistics Seminar
DESCRIPTION:Location: 2203 SAS Hall\, NC State Main Campus\n\nSpeaker: Dr. Vadim Zipunnikov\, Johns Hopkins Bloomberg School of Public Health\n\nTitle: Developing more sensitive endpoints by leveraging novel statistical methods for Digital\nHealth Technologies (DHTs) data\n\nAbstract: Digital Health Technologies (DHT) are now used to continuously track physical\nactivity and sleep in many clinical studies. This DHT data provides tremendous opportunities to\ndevelop novel more sensitive clinical trial endpoints. There is\, however\, a large gap between the\ncomplexity of DHT data and statistical methodology for fully leveraging the potential of DHT.\nThis talk will discuss recent developments of novel DHT-centric statistical methods that can\nprovide more sensitive endpoints by extracting and fusing together information from temporal\,\ndistributional\, and time-series aspects of DHT data.\n\nBio: Dr. Vadim Zipunnikov is an Associate Professor of Biostatistics at Johns Hopkins Bloomberg School of Public Health. He co-leads the Wearable and Implantable Technology (WIT) group at Johns Hopkins University and serves as the Biostatistics Director of the NIMH-funded Motor Activity Research Consortium for Health (mMARCH)\, overseeing collection and analysis of large-scale digital health data across several clinical sites globally. Dr. Zipunnikov’s research focuses on developing advanced statistical methods for analyzing multimodal digital health data (DHT) from wearables and smartphones\, including accelerometry\, heart rate\, glucose monitors\, and ecological momentary assessment (EMA). In his work\, Dr. Zipunnikov collaborates with the Food and Drug Administration on digital analytics for drug development\, having developed a two-year course for the FDA’s Office of Biostatistics to equip drug reviewers with essential skills in DHT analytics. His research has garnered press coverage in NIH Research Matters\, NIH Directors’ blog\, TIME\, Washington Post\, Wall Street Journal\, CNN\, and BBC Radio. He has mentored 14 PhD students and 4 postdoctoral fellows and has authored over 100 peer-reviewed publications on digital biomarkers of physical/motor activity\, sleep\, and circadian rhythmicity in neurological and mental health disorders\, including Alzheimer’s Disease\, Multiple Sclerosis\, and Bipolar Disorder. He is known for his collaborative approach\, mentoring emerging researchers\, developing novel DHT-centric methods\, and his smile.
URL:https://statistics.sciences.ncsu.edu/event/statistics-seminar-41/
LOCATION:2203 SAS Hall
CATEGORIES:College of Sciences Calendar,Department,Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=America/New_York:20251031T110000
DTEND;TZID=America/New_York:20251031T120000
DTSTAMP:20260504T094058
CREATED:20251020T173731Z
LAST-MODIFIED:20251020T174546Z
UID:28864-1761908400-1761912000@statistics.sciences.ncsu.edu
SUMMARY:Statistics Seminar
DESCRIPTION:Location: 2203 SAS Hall\, NC State Main Campus \nTitle: Synthetic Populations\, Personas and Agents \nPresenter: Georgiy Bobashev\, Ph.D. \nAbstract: Many experiments and estimate are not feasible or unethical to conduct with real people but possible in silico with synthetic individuals. I will present the construction\, and the use of geospatially explicit and statistically accurate person and household data\, which allow researchers to study community-and neighborhood-level effects\, design and test hypotheses that would not be possible without synthetic data. I will present the workflow for generating spatially explicit household- and individual-level synthetic populations for the United States representing 330 million individuals. Synthetic population could be used to probabilistically link multiple datasets for a specific purpose. There are statistical challenges with calibration and validation of these datasets. Agent-based models use these synthetic populations to study policy implications\, disease spread\, drug using behaviors\, etc. The use of synthetics individuals is now broadly expanding into health\, economic\, defense and other areas. With the developments of AI\, AI agents are taking over certain functions\, which creates new challenges in the development\, calibration and validation of these synthetic individuals.
URL:https://statistics.sciences.ncsu.edu/event/statistics-seminar-42/
LOCATION:2203 SAS Hall
CATEGORIES:College of Sciences Calendar,Department,Seminars
END:VEVENT
END:VCALENDAR