BEGIN:VCALENDAR VERSION:2.0 PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4// BEGIN:VEVENT UID:20251120T181454EST-7470omloVc@132.216.98.100 DTSTAMP:20251120T231454Z DESCRIPTION:Title: BdryGP: a boundary-integrated Gaussian process model for computer code emulation\n\nAbstract:\n\nWith advances in mathematical mod eling and computational methods\, complex phenomena (e.g.\, universe forma tions\, rocket propulsion) can now be reliably simulated via computer code . This code solves a complicated system of equations representing the unde rlying science of the problem. Such simulations can be very time-intensive \, requiring months of computation for a single run. Gaussian processes (G Ps) are widely used as predictive models for “emulating” this expensive co mputer code. Yet with limited training data on a high-dimensional paramete r space\, such models can suffer from poor predictive performance and phys ical interpretability. Fortunately\, in many physical applications\, there is additional boundary information on the code beforehand\, either from g overning physics or scientific knowledge. We propose a new BdryGP model wh ich incorporates such boundary information for prediction. We show that Bd ryGP not only enjoys improved convergence rates over standard GP models wh ich do not incorporate boundaries\, but is also more resistant to the ``cu rse-of-dimensionality” in nonparametric regression. We then demonstrate th e improved predictive performance and posterior contraction of the BdryGP model on several test problems in the literature.\n\nThis talk will featur e joint work of Liang Ding\, Simon Mak\, C. F. Jeff Wu.\n\n\n Speaker\n\n\n Simon Mak is an Assistant Professor in the Department of Statistical Scien ce at Duke University. Prior to Duke\, he was a postdoctoral fellow at the Stewart School of Industrial & Systems Engineering at Georgia Tech. He ha s a Ph.D. in Industrial Engineering and a M.Sc. in Statistics from Georgia Tech (2018)\, and a B.Sc. in Statistics and Actuarial Science from Simon Fraser University (2013).\n\nHis research focuses on developing methodolog ies with theoretical guarantees for big data reduction and small data anal ytics\, and applying these methods within an algorithmic workflow to solve real-world problems. His broad research interests are in engineering stat istics\, machine learning and experimental design\, with applications to m echanical engineering\, biomedical engineering\, financial engineering\, g enetics\, and signal processing.\n\nZoom Link\n\n \n\nMeeting ID: 924 5390 4989\n\nPasscode: 690084\n\n \n\n \n\n \n\n \n\n \n\n \n\n \n\n \n\n \n\n  \n\n \n\n \n DTSTART:20200918T193000Z DTEND:20200918T203000Z SUMMARY:Simon Mak (Duke University) URL:/mathstat/channels/event/simon-mak-duke-university -324550 END:VEVENT END:VCALENDAR