BEGIN:VCALENDAR VERSION:2.0 PRODID:-//132.216.98.100//NONSGML kigkonsult.se iCalcreator 2.20.4// BEGIN:VEVENT UID:20251029T140728EDT-5363PXNTLc@132.216.98.100 DTSTAMP:20251029T180728Z DESCRIPTION:Title: Regularized Fine-Tuning for Representation Multi-Task Le arning: Adaptivity\, Minimaxity\, and RobustnessRegularized Fine-Tuning fo r Representation Multi-Task Learning: Adaptivity\, Minimaxity\, and Robust ness.\n\nAbstract:\n\nWe study multi-task linear regression for a collecti on of tasks that share a latent\, low-dimensional structure. Each task’s r egression vector belongs to a subspace whose dimension\, denoted intrinsic dimension\, is much smaller than the ambient dimension. Unlike classical analyses that assume an identical subspace for every task\, we allow each task’s subspace to drift from a single reference subspace by a controllabl e similarity radius\, and we permit an unknown fraction of tasks to be out liers that violate the shared-structure assumption altogether. Our contrib utions are threefold. First\, adaptivity: we design a penalized empirical- risk algorithm and a spectral method.  Both algorithms automatically adjus t to the unknown similarity radius and to the proportion of outliers. Seco nd\, minimaxity: we prove information-theoretic lower bounds on the best a chievable prediction risk over this problem class and show that both algor ithms attain these bounds up to constant factors\; when no outliers are pr esent\, the spectral method is exactly minimax-optimal. Third\, robustness : for every choice of similarity radius and outlier proportion\, the propo sed estimators never incur larger expected prediction error than independe nt single-task regression\, while delivering strict improvements whenever tasks are even moderately similar and outliers are sparse. Additionally\, we introduce a thresholding algorithm to adapt to an unknown intrinsic dim ension. We conduct extensive numerical experiments to validate our theoret ical findings.\n\nSpeaker\n\nYang Feng is a Professor of Biostatistics in the School of Global Public Health at New York University\, where he is al so affiliated with the Center for Data Science. He earned his Ph.D. in Ope rations Research from Princeton University in 2010. His research centers o n the theoretical and methodological foundations of machine learning\, hig h-dimensional statistics\, network models\, and nonparametric statistics\, with applications in Alzheimer’s disease prognosis\, cancer subtype class ification\, genomics\, electronic health records\, and biomedical imaging\ , enabling more accurate models for risk assessment and clinical decision- making. His work has been supported by grants from the National Institutes of Health and the National Science Foundation (NSF)\, including the NSF C AREER Award. He currently serves as Associate Editor for several leading j ournals\, including the Journal of the American Statistical Association (J ASA)\, the Journal of Business & Economic Statistics\, the Journal of Comp utational & Graphical Statistics\, and the Annals of Applied Statistics. I n addition\, he will serve as Review Editor for JASA and The American Stat istician from 2026 to 2028. His professional recognitions include being na med a Fellow of the American Statistical Association and the Institute of Mathematical Statistics\, as well as an elected member of the Internationa l Statistical Institute.\n DTSTART:20251024T203000Z DTEND:20251024T213000Z LOCATION:Room 1104\, ºÚÁÏÍø College 2001\, CA\, QC\, Montreal\, H3A 1G1\, 2 001\, avenue ºÚÁÏÍø College SUMMARY:Yang Feng (New York University) URL:/mathstat/channels/event/yang-feng-new-york-univer sity-368546 END:VEVENT END:VCALENDAR